from customer_retention.analysis.notebook_progress import track_and_export_previous

track_and_export_previous("05_relationship_analysis.ipynb")

import numpy as np
import pandas as pd
import plotly.graph_objects as go
import yaml
from plotly.subplots import make_subplots

from customer_retention.analysis.auto_explorer import ExplorationFindings, ExplorationManager, RecommendationRegistry
from customer_retention.analysis.visualization import ChartBuilder, display_figure
from customer_retention.core.config.column_config import ColumnType
from customer_retention.core.config.experiments import (
    FINDINGS_DIR,
)
from customer_retention.core.utils.leakage import detect_leaking_features
from customer_retention.stages.profiling import RecommendationCategory, RelationshipRecommender

from pathlib import Path

from customer_retention.analysis.auto_explorer import load_notebook_findings

FINDINGS_PATH, _namespace, dataset_name = load_notebook_findings("05_relationship_analysis.ipynb")
print(f"Using: {FINDINGS_PATH}")

RECOMMENDATIONS_PATH = FINDINGS_PATH.replace("_findings.yaml", "_recommendations.yaml")

findings = ExplorationFindings.load(FINDINGS_PATH)

# Load data - handle aggregated vs standard paths
from customer_retention.analysis.auto_explorer.active_dataset_store import load_active_dataset
from customer_retention.stages.temporal import TEMPORAL_METADATA_COLS

if "_aggregated" in FINDINGS_PATH:
    source_path = Path(findings.source_path)
    if not source_path.is_absolute():
        if str(source_path).startswith("experiments"):
            source_path = Path("..") / source_path
        else:
            source_path = FINDINGS_DIR / source_path.name
    if source_path.is_dir():
        from customer_retention.integrations.adapters.factory import get_delta
        df = get_delta(force_local=True).read(str(source_path))
    elif source_path.is_file():
        df = pd.read_parquet(source_path)
    else:
        df = load_active_dataset(_namespace, dataset_name)
    data_source = f"aggregated:{source_path.name}"
else:
    df = load_active_dataset(_namespace, dataset_name)
    data_source = dataset_name

_df_cols = set(df.columns)
findings.columns = {k: v for k, v in findings.columns.items() if k in _df_cols}
if findings.target_column and findings.target_column not in _df_cols:
    findings.target_column = None

charts = ChartBuilder()

if Path(RECOMMENDATIONS_PATH).exists():
    with open(RECOMMENDATIONS_PATH, "r") as f:
        registry = RecommendationRegistry.from_dict(yaml.safe_load(f))
    print(f"Loaded existing recommendations: {len(registry.all_recommendations)} total")
else:
    registry = RecommendationRegistry()
    registry.init_bronze(findings.source_path)
    _entity_col = (findings.time_series_metadata.entity_column
                   if findings.time_series_metadata else None)
    registry.init_silver(_entity_col or "entity_id")
    registry.init_gold(findings.target_column or "target")
    print("Initialized new recommendation registry")

print(f"\nLoaded {len(df):,} rows from: {data_source}")

Using: /Users/Vital/python/CustomerRetention/experiments/runs/retail-e7471284/datasets/customer_retention_retail/findings/customer_retention_retail_aggregated_findings.yaml

Loaded existing recommendations: 177 total

Loaded 30,770 rows from: aggregated:customer_retention_retail_aggregated

EXCLUDE_LEAKING_FEATURES = []

_check_cols = [
    name for name, col in findings.columns.items()
    if col.inferred_type in [ColumnType.NUMERIC_CONTINUOUS, ColumnType.NUMERIC_DISCRETE, ColumnType.BINARY]
    and name != findings.target_column
    and name not in TEMPORAL_METADATA_COLS
]

_auto_leakers = detect_leaking_features(df, _check_cols, findings.target_column)
_all_excluded = sorted(set(_auto_leakers) | set(EXCLUDE_LEAKING_FEATURES))

if _all_excluded:
    for _col in _all_excluded:
        findings.columns.pop(_col, None)
    df = df.drop(columns=[c for c in _all_excluded if c in df.columns])
    findings.excluded_leaking_features = _all_excluded

    _auto_only = [c for c in _auto_leakers if c not in EXCLUDE_LEAKING_FEATURES]
    _manual_only = [c for c in EXCLUDE_LEAKING_FEATURES if c not in _auto_leakers]
    print(f"Excluded {len(_all_excluded)} leaking feature(s):")
    if _auto_only:
        print(f"  Auto-detected: {', '.join(_auto_only)}")
    if _manual_only:
        print(f"  Manual: {', '.join(_manual_only)}")
    _both = [c for c in _all_excluded if c in _auto_leakers and c in EXCLUDE_LEAKING_FEATURES]
    if _both:
        print(f"  Both: {', '.join(_both)}")
else:
    print("No leaking features detected.")

Excluded 1 leaking feature(s):
  Auto-detected: ordfreq_beginning

numeric_cols = [
    name for name, col in findings.columns.items()
    if col.inferred_type in [ColumnType.NUMERIC_CONTINUOUS, ColumnType.NUMERIC_DISCRETE, ColumnType.TARGET]
    and name not in TEMPORAL_METADATA_COLS
]

if len(numeric_cols) >= 2:
    corr_matrix = df[numeric_cols].corr()
    fig = charts.heatmap(
        corr_matrix.values,
        x_labels=numeric_cols,
        y_labels=numeric_cols,
        title="Numeric Correlation Matrix"
    )
    display_figure(fig)
else:
    print("Not enough numeric columns for correlation analysis.")

high_corr_threshold = 0.7
high_corr_pairs = []

if len(numeric_cols) >= 2:
    corr_matrix = df[numeric_cols].corr()
    for i in range(len(numeric_cols)):
        for j in range(i+1, len(numeric_cols)):
            corr_val = corr_matrix.iloc[i, j]
            if abs(corr_val) >= high_corr_threshold:
                high_corr_pairs.append({
                    "Column 1": numeric_cols[i],
                    "Column 2": numeric_cols[j],
                    "Correlation": f"{corr_val:.3f}"
                })

if high_corr_pairs:
    print(f"High Correlation Pairs (|r| >= {high_corr_threshold}):")
    display(pd.DataFrame(high_corr_pairs))
    print("\nConsider removing one of each pair to reduce multicollinearity.")
else:
    print("No high correlation pairs detected.")

High Correlation Pairs (|r| >= 0.7):

Consider removing one of each pair to reduce multicollinearity.

# Feature Distributions by Retention Status
if findings.target_column and findings.target_column in df.columns:
    target = findings.target_column

    feature_cols = [
        name for name, col in findings.columns.items()
        if col.inferred_type in [ColumnType.NUMERIC_CONTINUOUS, ColumnType.NUMERIC_DISCRETE]
        and name != target
        and name not in TEMPORAL_METADATA_COLS
    ]

    if feature_cols:
        print("=" * 80)
        print(f"FEATURE DISTRIBUTIONS BY TARGET: {target}")
        print("=" * 80)

        # Calculate summary statistics by target
        summary_by_target = []
        for col in feature_cols:
            for target_val, label in [(0, "Churned"), (1, "Retained")]:
                subset = df[df[target] == target_val][col].dropna()
                if len(subset) > 0:
                    summary_by_target.append({
                        "Feature": col,
                        "Group": label,
                        "Count": len(subset),
                        "Mean": subset.mean(),
                        "Median": subset.median(),
                        "Std": subset.std()
                    })

        if summary_by_target:
            summary_df = pd.DataFrame(summary_by_target)

            # Display summary table
            print("\n📊 Summary Statistics by Retention Status:")
            display_summary = summary_df.pivot(index="Feature", columns="Group", values=["Mean", "Median"])
            display_summary.columns = [f"{stat} ({group})" for stat, group in display_summary.columns]
            display(display_summary.round(3))

        # Calculate effect size (Cohen's d) for each feature
        print("\n📈 Feature Importance Indicators (Effect Size - Cohen's d):")
        print("-" * 70)
        effect_sizes = []
        for col in feature_cols:
            churned = df[df[target] == 0][col].dropna()
            retained = df[df[target] == 1][col].dropna()

            if len(churned) > 0 and len(retained) > 0:
                # Cohen's d
                pooled_std = np.sqrt(((len(churned)-1)*churned.std()**2 + (len(retained)-1)*retained.std()**2) /
                                     (len(churned) + len(retained) - 2))
                if pooled_std > 0:
                    d = (retained.mean() - churned.mean()) / pooled_std
                else:
                    d = 0

                # Interpret effect size
                abs_d = abs(d)
                if abs_d >= 0.8:
                    interpretation = "Large effect"
                    emoji = "🔴"
                elif abs_d >= 0.5:
                    interpretation = "Medium effect"
                    emoji = "🟡"
                elif abs_d >= 0.2:
                    interpretation = "Small effect"
                    emoji = "🟢"
                else:
                    interpretation = "Negligible"
                    emoji = "⚪"

                effect_sizes.append({
                    "feature": col,
                    "cohens_d": d,
                    "abs_d": abs_d,
                    "interpretation": interpretation
                })

                direction = "↑ Higher in retained" if d > 0 else "↓ Lower in retained"
                print(f"  {emoji} {col}: d={d:+.3f} ({interpretation}) {direction}")

        # Sort by effect size for identifying important features
        if effect_sizes:
            effect_df = pd.DataFrame(effect_sizes).sort_values("abs_d", ascending=False)
            important_features = effect_df[effect_df["abs_d"] >= 0.2]["feature"].tolist()
            if important_features:
                print(f"\n⭐ Features with notable effect (|d| ≥ 0.2): {', '.join(important_features)}")
        else:
            print("  No effect sizes could be calculated (insufficient data in one or both groups)")
    else:
        print("No numeric feature columns found for distribution analysis.")
else:
    print("Target column not available.")

================================================================================
FEATURE DISTRIBUTIONS BY TARGET: retained
================================================================================

📊 Summary Statistics by Retention Status:

📈 Feature Importance Indicators (Effect Size - Cohen's d):
----------------------------------------------------------------------
  ⚪ event_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained
  🔴 esent_sum_all_time: d=+2.551 (Large effect) ↑ Higher in retained
  🔴 esent_mean_all_time: d=+2.551 (Large effect) ↑ Higher in retained
  🔴 esent_max_all_time: d=+2.551 (Large effect) ↑ Higher in retained
  ⚪ esent_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained
  ⚪ eopenrate_sum_all_time: d=+0.187 (Negligible) ↑ Higher in retained
  ⚪ eopenrate_mean_all_time: d=+0.187 (Negligible) ↑ Higher in retained
  ⚪ eopenrate_max_all_time: d=+0.187 (Negligible) ↑ Higher in retained
  ⚪ eopenrate_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained
  ⚪ eclickrate_sum_all_time: d=+0.107 (Negligible) ↑ Higher in retained
  ⚪ eclickrate_mean_all_time: d=+0.107 (Negligible) ↑ Higher in retained
  ⚪ eclickrate_max_all_time: d=+0.107 (Negligible) ↑ Higher in retained

  ⚪ eclickrate_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained

  ⚪ avgorder_sum_all_time: d=+0.010 (Negligible) ↑ Higher in retained
  ⚪ avgorder_mean_all_time: d=+0.010 (Negligible) ↑ Higher in retained
  ⚪ avgorder_max_all_time: d=+0.010 (Negligible) ↑ Higher in retained
  ⚪ avgorder_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained
  ⚪ ordfreq_sum_all_time: d=+0.026 (Negligible) ↑ Higher in retained
  ⚪ ordfreq_mean_all_time: d=+0.026 (Negligible) ↑ Higher in retained
  ⚪ ordfreq_max_all_time: d=+0.026 (Negligible) ↑ Higher in retained
  ⚪ ordfreq_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained
  🟢 paperless_sum_all_time: d=+0.454 (Small effect) ↑ Higher in retained
  ⚪ paperless_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained
  ⚪ refill_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained
  ⚪ doorstep_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained

  ⚪ created_delta_hours_sum_all_time: d=+0.003 (Negligible) ↑ Higher in retained

  ⚪ created_delta_hours_mean_all_time: d=+0.009 (Negligible) ↑ Higher in retained
  ⚪ created_delta_hours_max_all_time: d=+0.009 (Negligible) ↑ Higher in retained
  ⚪ created_delta_hours_count_all_time: d=+0.113 (Negligible) ↑ Higher in retained
  ⚪ created_hour_sum_all_time: d=+0.000 (Negligible) ↓ Lower in retained
  ⚪ created_hour_mean_all_time: d=+0.000 (Negligible) ↓ Lower in retained
  ⚪ created_hour_max_all_time: d=+0.000 (Negligible) ↓ Lower in retained
  ⚪ created_hour_count_all_time: d=+0.119 (Negligible) ↑ Higher in retained
  ⚪ created_dow_sum_all_time: d=+0.042 (Negligible) ↑ Higher in retained
  ⚪ created_dow_mean_all_time: d=+0.026 (Negligible) ↑ Higher in retained
  ⚪ created_dow_max_all_time: d=+0.026 (Negligible) ↑ Higher in retained
  ⚪ created_dow_count_all_time: d=+0.119 (Negligible) ↑ Higher in retained
  ⚪ created_is_weekend_count_all_time: d=+0.119 (Negligible) ↑ Higher in retained
  ⚪ firstorder_delta_hours_sum_all_time: d=-0.040 (Negligible) ↓ Lower in retained

  ⚪ firstorder_delta_hours_mean_all_time: d=-0.041 (Negligible) ↓ Lower in retained

  ⚪ firstorder_delta_hours_max_all_time: d=-0.041 (Negligible) ↓ Lower in retained
  ⚪ firstorder_delta_hours_count_all_time: d=-0.004 (Negligible) ↓ Lower in retained
  ⚪ firstorder_hour_sum_all_time: d=+0.000 (Negligible) ↓ Lower in retained
  ⚪ firstorder_hour_mean_all_time: d=+0.000 (Negligible) ↓ Lower in retained
  ⚪ firstorder_hour_max_all_time: d=+0.000 (Negligible) ↓ Lower in retained
  ⚪ firstorder_hour_count_all_time: d=-0.004 (Negligible) ↓ Lower in retained
  ⚪ firstorder_dow_sum_all_time: d=+0.054 (Negligible) ↑ Higher in retained
  ⚪ firstorder_dow_mean_all_time: d=+0.054 (Negligible) ↑ Higher in retained
  ⚪ firstorder_dow_max_all_time: d=+0.054 (Negligible) ↑ Higher in retained
  ⚪ firstorder_dow_count_all_time: d=-0.004 (Negligible) ↓ Lower in retained
  ⚪ firstorder_is_weekend_mean_all_time: d=+0.059 (Negligible) ↑ Higher in retained
  ⚪ firstorder_is_weekend_count_all_time: d=+0.000 (Negligible) ↑ Higher in retained
  ⚪ days_since_created_sum_all_time: d=-0.003 (Negligible) ↓ Lower in retained

  ⚪ days_since_created_mean_all_time: d=-0.009 (Negligible) ↓ Lower in retained

# Box Plots: Visual comparison of distributions
if findings.target_column and findings.target_column in df.columns:
    target = findings.target_column

    feature_cols = [
        name for name, col in findings.columns.items()
        if col.inferred_type in [ColumnType.NUMERIC_CONTINUOUS, ColumnType.NUMERIC_DISCRETE]
        and name != target
        and name not in TEMPORAL_METADATA_COLS
    ]

    if feature_cols:
        # Create box plots - one subplot per feature for better control
        n_features = min(len(feature_cols), 6)

        fig = make_subplots(
            rows=1, cols=n_features,
            subplot_titles=feature_cols[:n_features],
            horizontal_spacing=0.05
        )

        for i, col in enumerate(feature_cols[:n_features]):
            col_num = i + 1

            # Retained (1) - Green
            retained_data = df[df[target] == 1][col].dropna()
            fig.add_trace(
                go.Box(
                    y=retained_data,
                    name='Retained',
                    fillcolor='rgba(46, 204, 113, 0.7)',
                    line=dict(color='#1e8449', width=2),
                    marker=dict(
                        color='rgba(46, 204, 113, 0.5)',  # Light green outliers
                        size=5,
                        line=dict(color='#1e8449', width=1)
                    ),
                    boxpoints='outliers',
                    width=0.35,
                    showlegend=(i == 0),
                    legendgroup='retained',
                    offsetgroup='retained'
                ),
                row=1, col=col_num
            )

            # Churned (0) - Red
            churned_data = df[df[target] == 0][col].dropna()
            fig.add_trace(
                go.Box(
                    y=churned_data,
                    name='Churned',
                    fillcolor='rgba(231, 76, 60, 0.7)',
                    line=dict(color='#922b21', width=2),
                    marker=dict(
                        color='rgba(231, 76, 60, 0.5)',  # Light red outliers
                        size=5,
                        line=dict(color='#922b21', width=1)
                    ),
                    boxpoints='outliers',
                    width=0.35,
                    showlegend=(i == 0),
                    legendgroup='churned',
                    offsetgroup='churned'
                ),
                row=1, col=col_num
            )

        fig.update_layout(
            height=450,
            title_text="Feature Distributions: Retained (Green) vs Churned (Red)",
            template='plotly_white',
            showlegend=True,
            legend=dict(orientation="h", yanchor="bottom", y=1.05, xanchor="center", x=0.5),
            boxmode='group',
            boxgap=0.3,
            boxgroupgap=0.1
        )

        # Center the boxes by removing x-axis tick labels (title is above each subplot)
        fig.update_xaxes(showticklabels=False)

        display_figure(fig)

        # Print mean comparison
        print("\n📊 MEAN COMPARISON BY RETENTION STATUS:")
        print("-" * 70)
        for col in feature_cols[:n_features]:
            retained_mean = df[df[target] == 1][col].mean()
            churned_mean = df[df[target] == 0][col].mean()
            diff_pct = ((retained_mean - churned_mean) / churned_mean * 100) if churned_mean != 0 else 0
            print(f"  {col}:")
            print(f"     Retained: {retained_mean:.2f}  |  Churned: {churned_mean:.2f}  |  Diff: {diff_pct:+.1f}%")

📊 MEAN COMPARISON BY RETENTION STATUS:
----------------------------------------------------------------------
  event_count_all_time:
     Retained: 1.00  |  Churned: 1.00  |  Diff: +0.0%
  esent_sum_all_time:
     Retained: 34.25  |  Churned: 4.49  |  Diff: +662.8%
  esent_mean_all_time:
     Retained: 34.25  |  Churned: 4.49  |  Diff: +662.8%
  esent_max_all_time:
     Retained: 34.25  |  Churned: 4.49  |  Diff: +662.7%
  esent_count_all_time:
     Retained: 1.00  |  Churned: 1.00  |  Diff: +0.0%
  eopenrate_sum_all_time:
     Retained: 26.69  |  Churned: 21.18  |  Diff: +26.0%

if findings.target_column and findings.target_column in df.columns:
    target = findings.target_column
    feature_cols = [
        name for name, col in findings.columns.items()
        if col.inferred_type in [ColumnType.NUMERIC_CONTINUOUS, ColumnType.NUMERIC_DISCRETE]
        and name != target
        and name not in TEMPORAL_METADATA_COLS
    ]

    if feature_cols:
        correlations = []
        for col in feature_cols:
            corr = df[[col, target]].corr().iloc[0, 1]
            correlations.append({"Feature": col, "Correlation": corr})

        corr_df = pd.DataFrame(correlations).sort_values("Correlation", key=abs, ascending=False)

        fig = charts.bar_chart(
            corr_df["Feature"].tolist(),
            corr_df["Correlation"].tolist(),
            title=f"Feature Correlations with {target}"
        )
        display_figure(fig)
else:
    print("Target column not available for correlation analysis.")

from customer_retention.stages.profiling import CategoricalTargetAnalyzer

if findings.target_column:
    target = findings.target_column
    overall_retention = df[target].mean()

    categorical_cols = [
        name for name, col in findings.columns.items()
        if col.inferred_type in [ColumnType.CATEGORICAL_NOMINAL, ColumnType.CATEGORICAL_ORDINAL]
        and name not in TEMPORAL_METADATA_COLS
    ]

    print("=" * 80)
    print("CATEGORICAL FEATURE ANALYSIS")
    print("=" * 80)
    print(f"Overall retention rate: {overall_retention:.1%}")

    if categorical_cols:
        # Use framework analyzer for summary
        cat_analyzer = CategoricalTargetAnalyzer(min_samples_per_category=10)
        summary_df = cat_analyzer.analyze_multiple(df, categorical_cols, target)

        print("\n📈 Categorical Feature Strength (Cramér's V):")
        print("-" * 60)
        for _, row in summary_df.iterrows():
            if row["cramers_v"] >= 0.3:
                strength = "Strong"
                emoji = "🔴"
            elif row["cramers_v"] >= 0.1:
                strength = "Moderate"
                emoji = "🟡"
            else:
                strength = "Weak"
                emoji = "🟢"
            sig = "***" if row["p_value"] < 0.001 else "**" if row["p_value"] < 0.01 else "*" if row["p_value"] < 0.05 else ""
            print(f"  {emoji} {row['feature']}: V={row['cramers_v']:.3f} ({strength}) {sig}")

        # Detailed analysis for each categorical feature
        for col_name in categorical_cols[:5]:
            result = cat_analyzer.analyze(df, col_name, target)

            print(f"\n{'='*60}")
            print(f"📊 {col_name.upper()}")
            print("="*60)

            # Display stats table
            if len(result.category_stats) > 0:
                display_stats = result.category_stats[['category', 'total_count', 'retention_rate', 'lift', 'pct_of_total']].copy()
                display_stats['retention_rate'] = display_stats['retention_rate'].apply(lambda x: f"{x:.1%}")
                display_stats['lift'] = display_stats['lift'].apply(lambda x: f"{x:.2f}x")
                display_stats['pct_of_total'] = display_stats['pct_of_total'].apply(lambda x: f"{x:.1%}")
                display_stats.columns = [col_name, 'Count', 'Retention Rate', 'Lift', '% of Data']
                display(display_stats)

                # Stacked bar chart
                cat_stats = result.category_stats
                categories = cat_stats['category'].tolist()
                retained_counts = cat_stats['retained_count'].tolist()
                churned_counts = cat_stats['churned_count'].tolist()

                fig = go.Figure()

                fig.add_trace(go.Bar(
                    name='Retained',
                    x=categories,
                    y=retained_counts,
                    marker_color='rgba(46, 204, 113, 0.8)',
                    text=[f"{r/(r+c)*100:.0f}%" for r, c in zip(retained_counts, churned_counts)],
                    textposition='inside',
                    textfont=dict(color='white', size=12)
                ))

                fig.add_trace(go.Bar(
                    name='Churned',
                    x=categories,
                    y=churned_counts,
                    marker_color='rgba(231, 76, 60, 0.8)',
                    text=[f"{c/(r+c)*100:.0f}%" for r, c in zip(retained_counts, churned_counts)],
                    textposition='inside',
                    textfont=dict(color='white', size=12)
                ))

                fig.update_layout(
                    barmode='stack',
                    title=f"Retention by {col_name}",
                    xaxis_title=col_name,
                    yaxis_title="Count",
                    template='plotly_white',
                    height=350,
                    legend=dict(orientation="h", yanchor="bottom", y=1.02, xanchor="center", x=0.5)
                )
                display_figure(fig)

                # Flag high-risk categories from framework result
                if result.high_risk_categories:
                    print("\n  ⚠️ High-risk categories (lift < 0.9x):")
                    for cat in result.high_risk_categories:
                        cat_row = cat_stats[cat_stats['category'] == cat].iloc[0]
                        print(f"     • {cat}: {cat_row['retention_rate']:.1%} retention ({cat_row['lift']:.2f}x lift)")
    else:
        print("\n  ℹ️ No categorical columns detected.")
else:
    print("No target column available for categorical analysis.")

================================================================================
CATEGORICAL FEATURE ANALYSIS
================================================================================
Overall retention rate: 79.5%

📈 Categorical Feature Strength (Cramér's V):
------------------------------------------------------------
  🟡 cohort_quarter: V=0.257 (Moderate) ***

============================================================
📊 COHORT_QUARTER
============================================================

  ⚠️ High-risk categories (lift < 0.9x):
     • 2016.0Q1.0: 71.2% retention (0.90x lift)
     • 2015.0Q3.0: 70.5% retention (0.89x lift)
     • 2012.0Q3.0: 69.9% retention (0.88x lift)
     • 2011.0Q4.0: 69.6% retention (0.88x lift)
     • 2015.0Q1.0: 69.6% retention (0.88x lift)
     • 2011.0Q1.0: 68.8% retention (0.87x lift)
     • 2015.0Q4.0: 67.7% retention (0.85x lift)
     • 2011.0Q3.0: 67.4% retention (0.85x lift)
     • 2015.0Q2.0: 67.2% retention (0.85x lift)
     • 2010.0Q4.0: 67.1% retention (0.84x lift)
     • 2011.0Q2.0: 65.9% retention (0.83x lift)
     • 2014.0Q2.0: 65.1% retention (0.82x lift)
     • 2008.0Q3.0: 64.7% retention (0.81x lift)
     • 2010.0Q3.0: 64.5% retention (0.81x lift)
     • 2009.0Q2.0: 60.5% retention (0.76x lift)
     • 2014.0Q4.0: 59.6% retention (0.75x lift)
     • 2010.0Q2.0: 58.4% retention (0.74x lift)
     • 2014.0Q3.0: 55.1% retention (0.69x lift)
     • 2010.0Q1.0: 54.9% retention (0.69x lift)
     • 2009.0Q3.0: 50.0% retention (0.63x lift)
     • 2009.0Q1.0: 50.0% retention (0.63x lift)
     • 2009.0Q4.0: 48.4% retention (0.61x lift)

top_numeric = numeric_cols[:4] if len(numeric_cols) > 4 else numeric_cols

if len(top_numeric) >= 2:
    fig = charts.scatter_matrix(
        df[top_numeric].sample(min(1000, len(df))),
        title="Scatter Plot Matrix (Sample)"
    )
    display_figure(fig)

from customer_retention.stages.profiling import TemporalTargetAnalyzer

datetime_cols = [
    name for name, col in findings.columns.items()
    if col.inferred_type == ColumnType.DATETIME
]

print("=" * 80)
print("DATETIME FEATURE ANALYSIS")
print("=" * 80)
print(f"Detected datetime columns: {datetime_cols}")

if datetime_cols and findings.target_column:
    target = findings.target_column
    overall_retention = df[target].mean()

    # Use framework analyzer
    temporal_analyzer = TemporalTargetAnalyzer(min_samples_per_period=10)

    for col_name in datetime_cols[:3]:
        result = temporal_analyzer.analyze(df, col_name, target)

        print(f"\n{'='*60}")
        print(f"📅 {col_name.upper()}")
        print("="*60)

        if result.n_valid_dates == 0:
            print("  No valid dates found")
            continue

        print(f"  Date range: {result.min_date} to {result.max_date}")
        print(f"  Valid dates: {result.n_valid_dates:,}")

        # 1. Retention by Year (from framework result)
        if len(result.yearly_stats) > 1:
            print(f"\n  📊 Retention by Year: Trend is {result.yearly_trend}")

            year_stats = result.yearly_stats

            fig = make_subplots(rows=1, cols=2, subplot_titles=["Retention Rate by Year", "Customer Count by Year"],
                               column_widths=[0.6, 0.4])

            fig.add_trace(
                go.Scatter(
                    x=year_stats['period'].astype(str),
                    y=year_stats['retention_rate'],
                    mode='lines+markers',
                    name='Retention Rate',
                    line=dict(color='#3498db', width=3),
                    marker=dict(size=10)
                ),
                row=1, col=1
            )
            fig.add_hline(y=overall_retention, line_dash="dash", line_color="gray",
                         annotation_text=f"Overall: {overall_retention:.1%}", row=1, col=1)

            fig.add_trace(
                go.Bar(
                    x=year_stats['period'].astype(str),
                    y=year_stats['count'],
                    name='Count',
                    marker_color='rgba(52, 152, 219, 0.6)'
                ),
                row=1, col=2
            )

            fig.update_layout(height=350, template='plotly_white', showlegend=False)
            fig.update_yaxes(tickformat='.0%', row=1, col=1)
            display_figure(fig)

        # 2. Retention by Month (from framework result)
        if len(result.monthly_stats) > 1:
            print("\n  📊 Retention by Month (Seasonality):")

            month_stats = result.monthly_stats
            colors = ['rgba(46, 204, 113, 0.7)' if r >= overall_retention else 'rgba(231, 76, 60, 0.7)'
                     for r in month_stats['retention_rate']]

            fig = go.Figure()
            fig.add_trace(go.Bar(
                x=month_stats['month_name'],
                y=month_stats['retention_rate'],
                marker_color=colors,
                text=[f"{r:.0%}" for r in month_stats['retention_rate']],
                textposition='outside'
            ))
            fig.add_hline(y=overall_retention, line_dash="dash", line_color="gray",
                         annotation_text=f"Overall: {overall_retention:.1%}")

            fig.update_layout(
                title=f"Monthly Retention Pattern ({col_name})",
                xaxis_title="Month",
                yaxis_title="Retention Rate",
                template='plotly_white',
                height=350,
                yaxis_tickformat='.0%'
            )
            display_figure(fig)

            # Seasonal insights from framework
            if result.seasonal_spread > 0.05:
                print(f"  📈 Seasonal spread: {result.seasonal_spread:.1%}")
                print(f"     Best month: {result.best_month}")
                print(f"     Worst month: {result.worst_month}")

        # 3. Retention by Day of Week (from framework result)
        if len(result.dow_stats) > 1:
            print("\n  📊 Retention by Day of Week:")

            dow_stats = result.dow_stats
            colors = ['rgba(46, 204, 113, 0.7)' if r >= overall_retention else 'rgba(231, 76, 60, 0.7)'
                     for r in dow_stats['retention_rate']]

            fig = go.Figure()
            fig.add_trace(go.Bar(
                x=dow_stats['day_name'],
                y=dow_stats['retention_rate'],
                marker_color=colors,
                text=[f"{r:.0%}" for r in dow_stats['retention_rate']],
                textposition='outside'
            ))
            fig.add_hline(y=overall_retention, line_dash="dash", line_color="gray")

            fig.update_layout(
                title=f"Day of Week Pattern ({col_name})",
                xaxis_title="Day of Week",
                yaxis_title="Retention Rate",
                template='plotly_white',
                height=300,
                yaxis_tickformat='.0%'
            )
            display_figure(fig)
else:
    if not datetime_cols:
        print("\n  ℹ️ No datetime columns detected in this dataset.")
        print("     Consider adding date parsing in notebook 01 if dates exist as strings.")
    else:
        print("\n  ℹ️ No target column available for retention analysis.")

================================================================================
DATETIME FEATURE ANALYSIS
================================================================================
Detected datetime columns: []

  ℹ️ No datetime columns detected in this dataset.
     Consider adding date parsing in notebook 01 if dates exist as strings.

# Generate comprehensive actionable recommendations
recommender = RelationshipRecommender()

# Gather columns by type
numeric_features = [
    name for name, col in findings.columns.items()
    if col.inferred_type in [ColumnType.NUMERIC_CONTINUOUS, ColumnType.NUMERIC_DISCRETE]
    and name != findings.target_column
    and name not in TEMPORAL_METADATA_COLS
]
categorical_features = [
    name for name, col in findings.columns.items()
    if col.inferred_type in [ColumnType.CATEGORICAL_NOMINAL, ColumnType.CATEGORICAL_ORDINAL]
    and name not in TEMPORAL_METADATA_COLS
]

# Run comprehensive analysis
analysis_summary = recommender.analyze(
    df,
    numeric_cols=numeric_features,
    categorical_cols=categorical_features,
    target_col=findings.target_column,
)

print("=" * 80)
print("ACTIONABLE RECOMMENDATIONS FROM RELATIONSHIP ANALYSIS")
print("=" * 80)

# Group recommendations by category
grouped_recs = analysis_summary.recommendations_by_category
high_priority = analysis_summary.high_priority_actions

if high_priority:
    print(f"\n🔴 HIGH PRIORITY ACTIONS ({len(high_priority)}):")
    print("-" * 60)
    for rec in high_priority:
        print(f"\n  📌 {rec.title}")
        print(f"     {rec.description}")
        print(f"     → Action: {rec.action}")
        if rec.affected_features:
            print(f"     → Features: {', '.join(rec.affected_features[:5])}")

# Persist recommendations to registry
for pair in analysis_summary.multicollinear_pairs:
    registry.add_gold_drop_multicollinear(
        column=pair["feature1"], correlated_with=pair["feature2"],
        correlation=pair["correlation"],
        rationale=f"High correlation ({pair['correlation']:.2f}) - consider dropping one",
        source_notebook="05_relationship_analysis"
    )

for predictor in analysis_summary.strong_predictors:
    registry.add_gold_prioritize_feature(
        column=predictor["feature"], effect_size=predictor["effect_size"],
        correlation=predictor["correlation"],
        rationale=f"Strong predictor with effect size {predictor['effect_size']:.2f}",
        source_notebook="05_relationship_analysis"
    )

for weak_col in analysis_summary.weak_predictors[:10]:
    registry.add_gold_drop_weak(
        column=weak_col, effect_size=0.0, correlation=0.0,
        rationale="Negligible predictive power",
        source_notebook="05_relationship_analysis"
    )

# Persist ratio feature recommendations
for rec in grouped_recs.get(RecommendationCategory.FEATURE_ENGINEERING, []):
    if "ratio" in rec.title.lower() and len(rec.affected_features) >= 2:
        registry.add_silver_ratio(
            column=f"{rec.affected_features[0]}_to_{rec.affected_features[1]}_ratio",
            numerator=rec.affected_features[0], denominator=rec.affected_features[1],
            rationale=rec.description, source_notebook="05_relationship_analysis"
        )
    elif "interaction" in rec.title.lower() and len(rec.affected_features) >= 2:
        for i, f1 in enumerate(rec.affected_features[:3]):
            for f2 in rec.affected_features[i+1:4]:
                registry.add_silver_interaction(
                    column=f"{f1}_x_{f2}", features=[f1, f2],
                    rationale=rec.description, source_notebook="05_relationship_analysis"
                )

# Store for findings metadata
findings.metadata["relationship_analysis"] = {
    "n_recommendations": len(analysis_summary.recommendations),
    "n_high_priority": len(high_priority),
    "strong_predictors": [p["feature"] for p in analysis_summary.strong_predictors],
    "weak_predictors": analysis_summary.weak_predictors[:5],
    "multicollinear_pairs": [(p["feature1"], p["feature2"]) for p in analysis_summary.multicollinear_pairs],
}

print(f"\n✅ Persisted {len(analysis_summary.multicollinear_pairs)} multicollinearity recommendations")
print(f"✅ Persisted {len(analysis_summary.strong_predictors)} strong predictor recommendations")
print(f"✅ Persisted {min(len(analysis_summary.weak_predictors), 10)} weak predictor recommendations")

================================================================================
ACTIONABLE RECOMMENDATIONS FROM RELATIONSHIP ANALYSIS
================================================================================

🔴 HIGH PRIORITY ACTIONS (750):
------------------------------------------------------------

  📌 Remove multicollinear feature
     event_count_all_time and esent_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, esent_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and eopenrate_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, eopenrate_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and eclickrate_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, eclickrate_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and avgorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, avgorder_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, ordfreq_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, paperless_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and refill_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, refill_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, doorstep_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     event_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: event_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     esent_sum_all_time and esent_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, esent_mean_all_time

  📌 Remove multicollinear feature
     esent_sum_all_time and esent_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, esent_max_all_time

  📌 Remove multicollinear feature
     esent_sum_all_time and lag0_esent_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, lag0_esent_sum

  📌 Remove multicollinear feature
     esent_sum_all_time and lag0_esent_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, lag0_esent_mean

  📌 Remove multicollinear feature
     esent_sum_all_time and lag0_esent_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, lag0_esent_max

  📌 Remove multicollinear feature
     esent_sum_all_time and esent_beginning are highly correlated (r=0.90)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, esent_beginning

  📌 Remove multicollinear feature
     esent_sum_all_time and esent_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, esent_end

  📌 Remove multicollinear feature
     esent_sum_all_time and esent_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, esent_vs_cohort_mean

  📌 Remove multicollinear feature
     esent_sum_all_time and esent_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, esent_vs_cohort_pct

  📌 Remove multicollinear feature
     esent_sum_all_time and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_sum_all_time, esent_cohort_zscore

  📌 Remove multicollinear feature
     esent_mean_all_time and esent_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, esent_max_all_time

  📌 Remove multicollinear feature
     esent_mean_all_time and lag0_esent_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, lag0_esent_sum

  📌 Remove multicollinear feature
     esent_mean_all_time and lag0_esent_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, lag0_esent_mean

  📌 Remove multicollinear feature
     esent_mean_all_time and lag0_esent_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, lag0_esent_max

  📌 Remove multicollinear feature
     esent_mean_all_time and esent_beginning are highly correlated (r=0.90)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, esent_beginning

  📌 Remove multicollinear feature
     esent_mean_all_time and esent_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, esent_end

  📌 Remove multicollinear feature
     esent_mean_all_time and esent_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, esent_vs_cohort_mean

  📌 Remove multicollinear feature
     esent_mean_all_time and esent_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, esent_vs_cohort_pct

  📌 Remove multicollinear feature
     esent_mean_all_time and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_mean_all_time, esent_cohort_zscore

  📌 Remove multicollinear feature
     esent_max_all_time and lag0_esent_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_max_all_time, lag0_esent_sum

  📌 Remove multicollinear feature
     esent_max_all_time and lag0_esent_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_max_all_time, lag0_esent_mean

  📌 Remove multicollinear feature
     esent_max_all_time and lag0_esent_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_max_all_time, lag0_esent_max

  📌 Remove multicollinear feature
     esent_max_all_time and esent_beginning are highly correlated (r=0.98)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_max_all_time, esent_beginning

  📌 Remove multicollinear feature
     esent_max_all_time and esent_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_max_all_time, esent_vs_cohort_mean

  📌 Remove multicollinear feature
     esent_max_all_time and esent_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_max_all_time, esent_vs_cohort_pct

  📌 Remove multicollinear feature
     esent_max_all_time and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_max_all_time, esent_cohort_zscore

  📌 Remove multicollinear feature
     esent_count_all_time and eopenrate_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, eopenrate_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and eclickrate_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, eclickrate_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and avgorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, avgorder_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, ordfreq_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, paperless_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and refill_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, refill_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, doorstep_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     esent_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and eopenrate_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, eopenrate_mean_all_time

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and eopenrate_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, eopenrate_max_all_time

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and lag0_eopenrate_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, lag0_eopenrate_sum

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and lag0_eopenrate_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, lag0_eopenrate_mean

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and lag0_eopenrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, lag0_eopenrate_max

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and eopenrate_beginning are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, eopenrate_beginning

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_sum_all_time and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_sum_all_time, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_mean_all_time and eopenrate_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_mean_all_time, eopenrate_max_all_time

  📌 Remove multicollinear feature
     eopenrate_mean_all_time and lag0_eopenrate_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_mean_all_time, lag0_eopenrate_sum

  📌 Remove multicollinear feature
     eopenrate_mean_all_time and lag0_eopenrate_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_mean_all_time, lag0_eopenrate_mean

  📌 Remove multicollinear feature
     eopenrate_mean_all_time and lag0_eopenrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_mean_all_time, lag0_eopenrate_max

  📌 Remove multicollinear feature
     eopenrate_mean_all_time and eopenrate_beginning are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_mean_all_time, eopenrate_beginning

  📌 Remove multicollinear feature
     eopenrate_mean_all_time and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_mean_all_time, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_mean_all_time and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_mean_all_time, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_mean_all_time and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_mean_all_time, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_max_all_time and lag0_eopenrate_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_max_all_time, lag0_eopenrate_sum

  📌 Remove multicollinear feature
     eopenrate_max_all_time and lag0_eopenrate_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_max_all_time, lag0_eopenrate_mean

  📌 Remove multicollinear feature
     eopenrate_max_all_time and lag0_eopenrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_max_all_time, lag0_eopenrate_max

  📌 Remove multicollinear feature
     eopenrate_max_all_time and eopenrate_beginning are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_max_all_time, eopenrate_beginning

  📌 Remove multicollinear feature
     eopenrate_max_all_time and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_max_all_time, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_max_all_time and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_max_all_time, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_max_all_time and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_max_all_time, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_count_all_time and eclickrate_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, eclickrate_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and avgorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, avgorder_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, ordfreq_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, paperless_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and refill_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, refill_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, doorstep_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     eopenrate_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and eclickrate_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, eclickrate_mean_all_time

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and eclickrate_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, eclickrate_max_all_time

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and lag0_eclickrate_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, lag0_eclickrate_sum

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and lag0_eclickrate_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, lag0_eclickrate_mean

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and lag0_eclickrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, lag0_eclickrate_max

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and eopenrate_end are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, eopenrate_end

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and eopenrate_trend_ratio are highly correlated (r=0.90)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and avgorder_beginning are highly correlated (r=-0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, avgorder_beginning

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and avgorder_end are highly correlated (r=-0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, avgorder_end

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, eclickrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eclickrate_sum_all_time and eclickrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_sum_all_time, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and eclickrate_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, eclickrate_max_all_time

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and lag0_eclickrate_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, lag0_eclickrate_sum

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and lag0_eclickrate_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, lag0_eclickrate_mean

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and lag0_eclickrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, lag0_eclickrate_max

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and eopenrate_end are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, eopenrate_end

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and eopenrate_trend_ratio are highly correlated (r=0.90)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and avgorder_beginning are highly correlated (r=-0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, avgorder_beginning

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and avgorder_end are highly correlated (r=-0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, avgorder_end

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, eclickrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eclickrate_mean_all_time and eclickrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_mean_all_time, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     eclickrate_max_all_time and lag0_eclickrate_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, lag0_eclickrate_sum

  📌 Remove multicollinear feature
     eclickrate_max_all_time and lag0_eclickrate_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, lag0_eclickrate_mean

  📌 Remove multicollinear feature
     eclickrate_max_all_time and lag0_eclickrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, lag0_eclickrate_max

  📌 Remove multicollinear feature
     eclickrate_max_all_time and eopenrate_end are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, eopenrate_end

  📌 Remove multicollinear feature
     eclickrate_max_all_time and eopenrate_trend_ratio are highly correlated (r=0.90)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     eclickrate_max_all_time and avgorder_beginning are highly correlated (r=-0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, avgorder_beginning

  📌 Remove multicollinear feature
     eclickrate_max_all_time and avgorder_end are highly correlated (r=-0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, avgorder_end

  📌 Remove multicollinear feature
     eclickrate_max_all_time and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, eclickrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eclickrate_max_all_time and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eclickrate_max_all_time and eclickrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_max_all_time, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     eclickrate_count_all_time and avgorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, avgorder_count_all_time

  📌 Remove multicollinear feature
     eclickrate_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, ordfreq_count_all_time

  📌 Remove multicollinear feature
     eclickrate_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, paperless_count_all_time

  📌 Remove multicollinear feature
     eclickrate_count_all_time and refill_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, refill_count_all_time

  📌 Remove multicollinear feature
     eclickrate_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, doorstep_count_all_time

  📌 Remove multicollinear feature
     eclickrate_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     eclickrate_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     eclickrate_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     eclickrate_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     avgorder_sum_all_time and avgorder_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, avgorder_mean_all_time

  📌 Remove multicollinear feature
     avgorder_sum_all_time and avgorder_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, avgorder_max_all_time

  📌 Remove multicollinear feature
     avgorder_sum_all_time and lag0_avgorder_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, lag0_avgorder_sum

  📌 Remove multicollinear feature
     avgorder_sum_all_time and lag0_avgorder_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, lag0_avgorder_mean

  📌 Remove multicollinear feature
     avgorder_sum_all_time and lag0_avgorder_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, lag0_avgorder_max

  📌 Remove multicollinear feature
     avgorder_sum_all_time and eopenrate_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, eopenrate_end

  📌 Remove multicollinear feature
     avgorder_sum_all_time and eopenrate_trend_ratio are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     avgorder_sum_all_time and avgorder_beginning are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, avgorder_beginning

  📌 Remove multicollinear feature
     avgorder_sum_all_time and avgorder_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, avgorder_end

  📌 Remove multicollinear feature
     avgorder_sum_all_time and avgorder_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     avgorder_sum_all_time and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     avgorder_sum_all_time and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_sum_all_time, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_mean_all_time and avgorder_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, avgorder_max_all_time

  📌 Remove multicollinear feature
     avgorder_mean_all_time and lag0_avgorder_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, lag0_avgorder_sum

  📌 Remove multicollinear feature
     avgorder_mean_all_time and lag0_avgorder_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, lag0_avgorder_mean

  📌 Remove multicollinear feature
     avgorder_mean_all_time and lag0_avgorder_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, lag0_avgorder_max

  📌 Remove multicollinear feature
     avgorder_mean_all_time and eopenrate_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, eopenrate_end

  📌 Remove multicollinear feature
     avgorder_mean_all_time and eopenrate_trend_ratio are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     avgorder_mean_all_time and avgorder_beginning are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, avgorder_beginning

  📌 Remove multicollinear feature
     avgorder_mean_all_time and avgorder_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, avgorder_end

  📌 Remove multicollinear feature
     avgorder_mean_all_time and avgorder_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     avgorder_mean_all_time and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     avgorder_mean_all_time and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_mean_all_time, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_max_all_time and lag0_avgorder_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, lag0_avgorder_sum

  📌 Remove multicollinear feature
     avgorder_max_all_time and lag0_avgorder_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, lag0_avgorder_mean

  📌 Remove multicollinear feature
     avgorder_max_all_time and lag0_avgorder_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, lag0_avgorder_max

  📌 Remove multicollinear feature
     avgorder_max_all_time and eopenrate_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, eopenrate_end

  📌 Remove multicollinear feature
     avgorder_max_all_time and eopenrate_trend_ratio are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     avgorder_max_all_time and avgorder_beginning are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, avgorder_beginning

  📌 Remove multicollinear feature
     avgorder_max_all_time and avgorder_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, avgorder_end

  📌 Remove multicollinear feature
     avgorder_max_all_time and avgorder_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     avgorder_max_all_time and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     avgorder_max_all_time and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_max_all_time, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_count_all_time, ordfreq_count_all_time

  📌 Remove multicollinear feature
     avgorder_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_count_all_time, paperless_count_all_time

  📌 Remove multicollinear feature
     avgorder_count_all_time and refill_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_count_all_time, refill_count_all_time

  📌 Remove multicollinear feature
     avgorder_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_count_all_time, doorstep_count_all_time

  📌 Remove multicollinear feature
     avgorder_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     avgorder_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     avgorder_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     avgorder_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     ordfreq_sum_all_time and ordfreq_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_sum_all_time, ordfreq_mean_all_time

  📌 Remove multicollinear feature
     ordfreq_sum_all_time and ordfreq_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_sum_all_time, ordfreq_max_all_time

  📌 Remove multicollinear feature
     ordfreq_sum_all_time and lag0_ordfreq_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_sum_all_time, lag0_ordfreq_sum

  📌 Remove multicollinear feature
     ordfreq_sum_all_time and lag0_ordfreq_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_sum_all_time, lag0_ordfreq_mean

  📌 Remove multicollinear feature
     ordfreq_sum_all_time and lag0_ordfreq_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_sum_all_time, lag0_ordfreq_max

  📌 Remove multicollinear feature
     ordfreq_sum_all_time and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_sum_all_time, ordfreq_vs_cohort_mean

  📌 Remove multicollinear feature
     ordfreq_sum_all_time and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_sum_all_time, ordfreq_vs_cohort_pct

  📌 Remove multicollinear feature
     ordfreq_sum_all_time and ordfreq_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_sum_all_time, ordfreq_cohort_zscore

  📌 Remove multicollinear feature
     ordfreq_mean_all_time and ordfreq_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_mean_all_time, ordfreq_max_all_time

  📌 Remove multicollinear feature
     ordfreq_mean_all_time and lag0_ordfreq_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_mean_all_time, lag0_ordfreq_sum

  📌 Remove multicollinear feature
     ordfreq_mean_all_time and lag0_ordfreq_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_mean_all_time, lag0_ordfreq_mean

  📌 Remove multicollinear feature
     ordfreq_mean_all_time and lag0_ordfreq_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_mean_all_time, lag0_ordfreq_max

  📌 Remove multicollinear feature
     ordfreq_mean_all_time and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_mean_all_time, ordfreq_vs_cohort_mean

  📌 Remove multicollinear feature
     ordfreq_mean_all_time and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_mean_all_time, ordfreq_vs_cohort_pct

  📌 Remove multicollinear feature
     ordfreq_mean_all_time and ordfreq_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_mean_all_time, ordfreq_cohort_zscore

  📌 Remove multicollinear feature
     ordfreq_max_all_time and lag0_ordfreq_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_max_all_time, lag0_ordfreq_sum

  📌 Remove multicollinear feature
     ordfreq_max_all_time and lag0_ordfreq_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_max_all_time, lag0_ordfreq_mean

  📌 Remove multicollinear feature
     ordfreq_max_all_time and lag0_ordfreq_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_max_all_time, lag0_ordfreq_max

  📌 Remove multicollinear feature
     ordfreq_max_all_time and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_max_all_time, ordfreq_vs_cohort_mean

  📌 Remove multicollinear feature
     ordfreq_max_all_time and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_max_all_time, ordfreq_vs_cohort_pct

  📌 Remove multicollinear feature
     ordfreq_max_all_time and ordfreq_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_max_all_time, ordfreq_cohort_zscore

  📌 Remove multicollinear feature
     ordfreq_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_count_all_time, paperless_count_all_time

  📌 Remove multicollinear feature
     ordfreq_count_all_time and refill_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_count_all_time, refill_count_all_time

  📌 Remove multicollinear feature
     ordfreq_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_count_all_time, doorstep_count_all_time

  📌 Remove multicollinear feature
     ordfreq_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     ordfreq_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     ordfreq_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     ordfreq_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     paperless_sum_all_time and esent_trend_ratio are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_sum_all_time, esent_trend_ratio

  📌 Remove multicollinear feature
     paperless_sum_all_time and avgorder_beginning are highly correlated (r=-0.92)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_sum_all_time, avgorder_beginning

  📌 Remove multicollinear feature
     paperless_sum_all_time and avgorder_end are highly correlated (r=-0.92)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_sum_all_time, avgorder_end

  📌 Remove multicollinear feature
     paperless_count_all_time and refill_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_count_all_time, refill_count_all_time

  📌 Remove multicollinear feature
     paperless_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_count_all_time, doorstep_count_all_time

  📌 Remove multicollinear feature
     paperless_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     paperless_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     paperless_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     paperless_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: paperless_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     refill_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: refill_count_all_time, doorstep_count_all_time

  📌 Remove multicollinear feature
     refill_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: refill_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     refill_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: refill_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     refill_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: refill_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     refill_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: refill_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     doorstep_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: doorstep_count_all_time, firstorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     doorstep_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: doorstep_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     doorstep_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: doorstep_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     doorstep_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: doorstep_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and created_delta_hours_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, created_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and created_delta_hours_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, created_delta_hours_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_since_created_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_since_created_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_since_created_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_since_created_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_since_created_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_since_created_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_until_created_sum_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_until_created_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_until_created_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_until_created_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_until_created_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     created_delta_hours_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_sum_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and created_delta_hours_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, created_delta_hours_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_since_created_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_since_created_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_since_created_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_since_created_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_since_created_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_since_created_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_until_created_sum_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_until_created_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_until_created_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_until_created_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_until_created_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     created_delta_hours_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_mean_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_since_created_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_since_created_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_since_created_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_since_created_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_since_created_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_since_created_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_until_created_sum_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_until_created_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_until_created_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_until_created_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_until_created_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     created_delta_hours_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_max_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and created_hour_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, created_hour_count_all_time

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and created_dow_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, created_dow_count_all_time

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and created_is_weekend_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, created_is_weekend_count_all_time

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and days_since_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, days_since_created_count_all_time

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and days_until_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, days_until_created_count_all_time

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, log1p_days_since_created_count_all_time

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and is_future_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, is_future_created_count_all_time

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     created_delta_hours_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     created_hour_count_all_time and created_dow_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_hour_count_all_time, created_dow_count_all_time

  📌 Remove multicollinear feature
     created_hour_count_all_time and created_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_hour_count_all_time, created_is_weekend_count_all_time

  📌 Remove multicollinear feature
     created_hour_count_all_time and days_since_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_hour_count_all_time, days_since_created_count_all_time

  📌 Remove multicollinear feature
     created_hour_count_all_time and days_until_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_hour_count_all_time, days_until_created_count_all_time

  📌 Remove multicollinear feature
     created_hour_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_hour_count_all_time, log1p_days_since_created_count_all_time

  📌 Remove multicollinear feature
     created_hour_count_all_time and is_future_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_hour_count_all_time, is_future_created_count_all_time

  📌 Remove multicollinear feature
     created_hour_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_hour_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     created_hour_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_hour_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     created_dow_sum_all_time and created_dow_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_sum_all_time, created_dow_mean_all_time

  📌 Remove multicollinear feature
     created_dow_sum_all_time and created_dow_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_sum_all_time, created_dow_max_all_time

  📌 Remove multicollinear feature
     created_dow_mean_all_time and created_dow_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_mean_all_time, created_dow_max_all_time

  📌 Remove multicollinear feature
     created_dow_count_all_time and created_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_count_all_time, created_is_weekend_count_all_time

  📌 Remove multicollinear feature
     created_dow_count_all_time and days_since_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_count_all_time, days_since_created_count_all_time

  📌 Remove multicollinear feature
     created_dow_count_all_time and days_until_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_count_all_time, days_until_created_count_all_time

  📌 Remove multicollinear feature
     created_dow_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_count_all_time, log1p_days_since_created_count_all_time

  📌 Remove multicollinear feature
     created_dow_count_all_time and is_future_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_count_all_time, is_future_created_count_all_time

  📌 Remove multicollinear feature
     created_dow_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     created_dow_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_dow_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     created_is_weekend_count_all_time and days_since_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_is_weekend_count_all_time, days_since_created_count_all_time

  📌 Remove multicollinear feature
     created_is_weekend_count_all_time and days_until_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_is_weekend_count_all_time, days_until_created_count_all_time

  📌 Remove multicollinear feature
     created_is_weekend_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_is_weekend_count_all_time, log1p_days_since_created_count_all_time

  📌 Remove multicollinear feature
     created_is_weekend_count_all_time and is_future_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_is_weekend_count_all_time, is_future_created_count_all_time

  📌 Remove multicollinear feature
     created_is_weekend_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_is_weekend_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     created_is_weekend_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_is_weekend_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     firstorder_delta_hours_sum_all_time and firstorder_delta_hours_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_delta_hours_sum_all_time, firstorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     firstorder_delta_hours_sum_all_time and firstorder_delta_hours_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_delta_hours_sum_all_time, firstorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     firstorder_delta_hours_sum_all_time and eopenrate_beginning are highly correlated (r=0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_delta_hours_sum_all_time, eopenrate_beginning

  📌 Remove multicollinear feature
     firstorder_delta_hours_mean_all_time and firstorder_delta_hours_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_delta_hours_mean_all_time, firstorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     firstorder_delta_hours_mean_all_time and eopenrate_beginning are highly correlated (r=0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_delta_hours_mean_all_time, eopenrate_beginning

  📌 Remove multicollinear feature
     firstorder_delta_hours_max_all_time and eopenrate_beginning are highly correlated (r=0.93)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_delta_hours_max_all_time, eopenrate_beginning

  📌 Remove multicollinear feature
     firstorder_delta_hours_count_all_time and firstorder_hour_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_delta_hours_count_all_time, firstorder_hour_count_all_time

  📌 Remove multicollinear feature
     firstorder_delta_hours_count_all_time and firstorder_dow_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_delta_hours_count_all_time, firstorder_dow_count_all_time

  📌 Remove multicollinear feature
     firstorder_hour_count_all_time and firstorder_dow_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_hour_count_all_time, firstorder_dow_count_all_time

  📌 Remove multicollinear feature
     firstorder_dow_sum_all_time and firstorder_dow_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_dow_sum_all_time, firstorder_dow_mean_all_time

  📌 Remove multicollinear feature
     firstorder_dow_sum_all_time and firstorder_dow_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_dow_sum_all_time, firstorder_dow_max_all_time

  📌 Remove multicollinear feature
     firstorder_dow_mean_all_time and firstorder_dow_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_dow_mean_all_time, firstorder_dow_max_all_time

  📌 Remove multicollinear feature
     firstorder_is_weekend_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_is_weekend_count_all_time, is_missing_created_count_all_time

  📌 Remove multicollinear feature
     firstorder_is_weekend_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_is_weekend_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     firstorder_is_weekend_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: firstorder_is_weekend_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_since_created_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_since_created_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_since_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_since_created_max_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_until_created_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_until_created_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_until_created_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_until_created_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_until_created_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_until_created_max_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_since_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_since_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_since_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_since_created_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_sum_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_since_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_since_created_max_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_until_created_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_until_created_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_until_created_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_until_created_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_until_created_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_until_created_max_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_since_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_since_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_since_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_since_created_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_mean_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_until_created_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_until_created_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_until_created_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_until_created_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_until_created_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_until_created_max_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_since_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_since_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_since_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and days_until_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_since_created_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_since_created_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_since_created_max_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_since_created_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_since_created_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_since_created_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_max_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_since_created_count_all_time and days_until_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_count_all_time, days_until_created_count_all_time

  📌 Remove multicollinear feature
     days_since_created_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_count_all_time, log1p_days_since_created_count_all_time

  📌 Remove multicollinear feature
     days_since_created_count_all_time and is_future_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_count_all_time, is_future_created_count_all_time

  📌 Remove multicollinear feature
     days_since_created_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     days_since_created_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_created_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and days_until_created_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, days_until_created_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, days_until_created_max_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_until_created_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_sum_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, days_until_created_max_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_until_created_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_mean_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_until_created_max_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, days_since_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_until_created_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_until_created_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_until_created_max_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_until_created_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_until_created_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_until_created_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_max_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_until_created_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_count_all_time, log1p_days_since_created_count_all_time

  📌 Remove multicollinear feature
     days_until_created_count_all_time and is_future_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_count_all_time, is_future_created_count_all_time

  📌 Remove multicollinear feature
     days_until_created_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     days_until_created_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_created_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     log1p_days_since_created_sum_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_created_sum_all_time, log1p_days_since_created_mean_all_time

  📌 Remove multicollinear feature
     log1p_days_since_created_sum_all_time and log1p_days_since_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_created_sum_all_time, log1p_days_since_created_max_all_time

  📌 Remove multicollinear feature
     log1p_days_since_created_mean_all_time and log1p_days_since_created_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_created_mean_all_time, log1p_days_since_created_max_all_time

  📌 Remove multicollinear feature
     log1p_days_since_created_count_all_time and is_future_created_count_all_time are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_created_count_all_time, is_future_created_count_all_time

  📌 Remove multicollinear feature
     log1p_days_since_created_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_created_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     log1p_days_since_created_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_created_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     is_missing_created_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: is_missing_created_count_all_time, is_missing_firstorder_count_all_time

  📌 Remove multicollinear feature
     is_missing_created_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: is_missing_created_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     is_future_created_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: is_future_created_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     is_future_created_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: is_future_created_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and days_since_firstorder_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and days_since_firstorder_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_since_firstorder_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_sum_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and days_since_firstorder_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_since_firstorder_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_mean_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, days_until_firstorder_sum_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and days_until_firstorder_max_all_time are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_since_firstorder_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_max_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_since_firstorder_count_all_time and days_until_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_count_all_time, days_until_firstorder_count_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_count_all_time and log1p_days_since_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_count_all_time, log1p_days_since_firstorder_count_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_count_all_time, lastorder_delta_hours_count_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_count_all_time, lastorder_hour_count_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_count_all_time, lastorder_dow_count_all_time

  📌 Remove multicollinear feature
     days_since_firstorder_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     days_since_firstorder_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_firstorder_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     days_until_firstorder_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_sum_all_time, days_until_firstorder_mean_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_sum_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_sum_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_until_firstorder_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_sum_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_until_firstorder_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_sum_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_until_firstorder_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_sum_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_until_firstorder_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_sum_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_until_firstorder_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_sum_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, days_until_firstorder_max_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_until_firstorder_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_mean_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, lastorder_delta_hours_sum_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.86)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and lag0_created_delta_hours_max are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     days_until_firstorder_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_max_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     days_until_firstorder_count_all_time and log1p_days_since_firstorder_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_count_all_time, log1p_days_since_firstorder_count_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_count_all_time, lastorder_delta_hours_count_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_count_all_time, lastorder_hour_count_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_count_all_time, lastorder_dow_count_all_time

  📌 Remove multicollinear feature
     days_until_firstorder_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     days_until_firstorder_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_until_firstorder_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     log1p_days_since_firstorder_sum_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_firstorder_sum_all_time, log1p_days_since_firstorder_mean_all_time

  📌 Remove multicollinear feature
     log1p_days_since_firstorder_sum_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_firstorder_sum_all_time, log1p_days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     log1p_days_since_firstorder_mean_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_firstorder_mean_all_time, log1p_days_since_firstorder_max_all_time

  📌 Remove multicollinear feature
     log1p_days_since_firstorder_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_firstorder_count_all_time, lastorder_delta_hours_count_all_time

  📌 Remove multicollinear feature
     log1p_days_since_firstorder_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_firstorder_count_all_time, lastorder_hour_count_all_time

  📌 Remove multicollinear feature
     log1p_days_since_firstorder_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_firstorder_count_all_time, lastorder_dow_count_all_time

  📌 Remove multicollinear feature
     log1p_days_since_firstorder_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_firstorder_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     log1p_days_since_firstorder_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: log1p_days_since_firstorder_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     is_missing_firstorder_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: is_missing_firstorder_count_all_time, lastorder_is_weekend_count_all_time

  📌 Remove multicollinear feature
     is_future_firstorder_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: is_future_firstorder_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     is_future_firstorder_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: is_future_firstorder_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, lastorder_delta_hours_mean_all_time

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and eopenrate_end are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, eopenrate_end

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and eopenrate_trend_ratio are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     lastorder_delta_hours_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_sum_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, lastorder_delta_hours_max_all_time

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and eopenrate_end are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, eopenrate_end

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and eopenrate_trend_ratio are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     lastorder_delta_hours_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_mean_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     lastorder_delta_hours_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_max_all_time, lag0_created_delta_hours_sum

  📌 Remove multicollinear feature
     lastorder_delta_hours_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_max_all_time, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     lastorder_delta_hours_max_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_max_all_time, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     lastorder_delta_hours_max_all_time and eopenrate_end are highly correlated (r=0.89)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_max_all_time, eopenrate_end

  📌 Remove multicollinear feature
     lastorder_delta_hours_max_all_time and eopenrate_trend_ratio are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_max_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lastorder_delta_hours_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_max_all_time, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     lastorder_delta_hours_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_max_all_time, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     lastorder_delta_hours_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.95)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_max_all_time, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     lastorder_delta_hours_count_all_time and lastorder_hour_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_count_all_time, lastorder_hour_count_all_time

  📌 Remove multicollinear feature
     lastorder_delta_hours_count_all_time and lastorder_dow_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_count_all_time, lastorder_dow_count_all_time

  📌 Remove multicollinear feature
     lastorder_delta_hours_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     lastorder_delta_hours_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_delta_hours_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lastorder_hour_count_all_time and lastorder_dow_count_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_hour_count_all_time, lastorder_dow_count_all_time

  📌 Remove multicollinear feature
     lastorder_hour_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_hour_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     lastorder_hour_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_hour_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lastorder_dow_sum_all_time and lastorder_dow_mean_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_sum_all_time, lastorder_dow_mean_all_time

  📌 Remove multicollinear feature
     lastorder_dow_sum_all_time and lastorder_dow_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_sum_all_time, lastorder_dow_max_all_time

  📌 Remove multicollinear feature
     lastorder_dow_sum_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_sum_all_time, eopenrate_end

  📌 Remove multicollinear feature
     lastorder_dow_sum_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_sum_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lastorder_dow_mean_all_time and lastorder_dow_max_all_time are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_mean_all_time, lastorder_dow_max_all_time

  📌 Remove multicollinear feature
     lastorder_dow_mean_all_time and eopenrate_end are highly correlated (r=-0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_mean_all_time, eopenrate_end

  📌 Remove multicollinear feature
     lastorder_dow_mean_all_time and eopenrate_trend_ratio are highly correlated (r=-0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_mean_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lastorder_dow_count_all_time and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_count_all_time, eopenrate_end

  📌 Remove multicollinear feature
     lastorder_dow_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lastorder_dow_count_all_time, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     days_since_first_event_x and cohort_year are highly correlated (r=-0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_first_event_x, cohort_year

  📌 Remove multicollinear feature
     days_since_first_event_x and avgorder_beginning are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_first_event_x, avgorder_beginning

  📌 Remove multicollinear feature
     days_since_first_event_x and avgorder_end are highly correlated (r=0.96)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: days_since_first_event_x, avgorder_end

  📌 Remove multicollinear feature
     cohort_year and avgorder_beginning are highly correlated (r=-0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: cohort_year, avgorder_beginning

  📌 Remove multicollinear feature
     cohort_year and avgorder_end are highly correlated (r=-0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: cohort_year, avgorder_end

  📌 Remove multicollinear feature
     lag0_esent_sum and lag0_esent_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_sum, lag0_esent_mean

  📌 Remove multicollinear feature
     lag0_esent_sum and lag0_esent_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_sum, lag0_esent_max

  📌 Remove multicollinear feature
     lag0_esent_sum and esent_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_sum, esent_end

  📌 Remove multicollinear feature
     lag0_esent_sum and esent_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_sum, esent_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_esent_sum and esent_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_sum, esent_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_esent_sum and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_sum, esent_cohort_zscore

  📌 Remove multicollinear feature
     lag0_esent_mean and lag0_esent_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_mean, lag0_esent_max

  📌 Remove multicollinear feature
     lag0_esent_mean and esent_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_mean, esent_end

  📌 Remove multicollinear feature
     lag0_esent_mean and esent_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_mean, esent_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_esent_mean and esent_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_mean, esent_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_esent_mean and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_mean, esent_cohort_zscore

  📌 Remove multicollinear feature
     lag0_esent_max and esent_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_max, esent_end

  📌 Remove multicollinear feature
     lag0_esent_max and esent_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_max, esent_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_esent_max and esent_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_max, esent_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_esent_max and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_esent_max, esent_cohort_zscore

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and lag0_eopenrate_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, lag0_eopenrate_mean

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and lag0_eopenrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, lag0_eopenrate_max

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and eopenrate_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, eopenrate_end

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and eopenrate_trend_ratio are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and avgorder_beginning are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, avgorder_beginning

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and avgorder_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, avgorder_end

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_eopenrate_sum and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_sum, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     lag0_eopenrate_mean and lag0_eopenrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_mean, lag0_eopenrate_max

  📌 Remove multicollinear feature
     lag0_eopenrate_mean and eopenrate_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_mean, eopenrate_end

  📌 Remove multicollinear feature
     lag0_eopenrate_mean and eopenrate_trend_ratio are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_mean, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_eopenrate_mean and avgorder_beginning are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_mean, avgorder_beginning

  📌 Remove multicollinear feature
     lag0_eopenrate_mean and avgorder_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_mean, avgorder_end

  📌 Remove multicollinear feature
     lag0_eopenrate_mean and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_mean, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_eopenrate_mean and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_mean, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_eopenrate_mean and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_mean, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     lag0_eopenrate_max and eopenrate_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_max, eopenrate_end

  📌 Remove multicollinear feature
     lag0_eopenrate_max and eopenrate_trend_ratio are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_max, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_eopenrate_max and avgorder_beginning are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_max, avgorder_beginning

  📌 Remove multicollinear feature
     lag0_eopenrate_max and avgorder_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_max, avgorder_end

  📌 Remove multicollinear feature
     lag0_eopenrate_max and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_max, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_eopenrate_max and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_max, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_eopenrate_max and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eopenrate_max, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     lag0_eclickrate_sum and lag0_eclickrate_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_sum, lag0_eclickrate_mean

  📌 Remove multicollinear feature
     lag0_eclickrate_sum and lag0_eclickrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_sum, lag0_eclickrate_max

  📌 Remove multicollinear feature
     lag0_eclickrate_sum and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_sum, eopenrate_end

  📌 Remove multicollinear feature
     lag0_eclickrate_sum and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_sum, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_eclickrate_sum and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_sum, eclickrate_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_eclickrate_sum and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_sum, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_eclickrate_sum and eclickrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_sum, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     lag0_eclickrate_mean and lag0_eclickrate_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_mean, lag0_eclickrate_max

  📌 Remove multicollinear feature
     lag0_eclickrate_mean and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_mean, eopenrate_end

  📌 Remove multicollinear feature
     lag0_eclickrate_mean and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_mean, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_eclickrate_mean and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_mean, eclickrate_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_eclickrate_mean and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_mean, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_eclickrate_mean and eclickrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_mean, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     lag0_eclickrate_max and eopenrate_end are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_max, eopenrate_end

  📌 Remove multicollinear feature
     lag0_eclickrate_max and eopenrate_trend_ratio are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_max, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_eclickrate_max and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_max, eclickrate_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_eclickrate_max and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_max, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_eclickrate_max and eclickrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_eclickrate_max, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     lag0_avgorder_sum and lag0_avgorder_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, lag0_avgorder_mean

  📌 Remove multicollinear feature
     lag0_avgorder_sum and lag0_avgorder_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, lag0_avgorder_max

  📌 Remove multicollinear feature
     lag0_avgorder_sum and eopenrate_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, eopenrate_end

  📌 Remove multicollinear feature
     lag0_avgorder_sum and eopenrate_trend_ratio are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_avgorder_sum and avgorder_beginning are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, avgorder_beginning

  📌 Remove multicollinear feature
     lag0_avgorder_sum and avgorder_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, avgorder_end

  📌 Remove multicollinear feature
     lag0_avgorder_sum and avgorder_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_avgorder_sum and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_avgorder_sum and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_sum, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     lag0_avgorder_mean and lag0_avgorder_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_mean, lag0_avgorder_max

  📌 Remove multicollinear feature
     lag0_avgorder_mean and eopenrate_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_mean, eopenrate_end

  📌 Remove multicollinear feature
     lag0_avgorder_mean and eopenrate_trend_ratio are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_mean, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_avgorder_mean and avgorder_beginning are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_mean, avgorder_beginning

  📌 Remove multicollinear feature
     lag0_avgorder_mean and avgorder_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_mean, avgorder_end

  📌 Remove multicollinear feature
     lag0_avgorder_mean and avgorder_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_mean, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_avgorder_mean and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_mean, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_avgorder_mean and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_mean, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     lag0_avgorder_max and eopenrate_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_max, eopenrate_end

  📌 Remove multicollinear feature
     lag0_avgorder_max and eopenrate_trend_ratio are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_max, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     lag0_avgorder_max and avgorder_beginning are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_max, avgorder_beginning

  📌 Remove multicollinear feature
     lag0_avgorder_max and avgorder_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_max, avgorder_end

  📌 Remove multicollinear feature
     lag0_avgorder_max and avgorder_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_max, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_avgorder_max and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_max, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_avgorder_max and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_avgorder_max, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     lag0_ordfreq_sum and lag0_ordfreq_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_sum, lag0_ordfreq_mean

  📌 Remove multicollinear feature
     lag0_ordfreq_sum and lag0_ordfreq_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_sum, lag0_ordfreq_max

  📌 Remove multicollinear feature
     lag0_ordfreq_sum and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_sum, ordfreq_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_ordfreq_sum and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_sum, ordfreq_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_ordfreq_sum and ordfreq_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_sum, ordfreq_cohort_zscore

  📌 Remove multicollinear feature
     lag0_ordfreq_mean and lag0_ordfreq_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_mean, lag0_ordfreq_max

  📌 Remove multicollinear feature
     lag0_ordfreq_mean and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_mean, ordfreq_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_ordfreq_mean and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_mean, ordfreq_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_ordfreq_mean and ordfreq_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_mean, ordfreq_cohort_zscore

  📌 Remove multicollinear feature
     lag0_ordfreq_max and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_max, ordfreq_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_ordfreq_max and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_max, ordfreq_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_ordfreq_max and ordfreq_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_ordfreq_max, ordfreq_cohort_zscore

  📌 Remove multicollinear feature
     lag0_created_delta_hours_sum and lag0_created_delta_hours_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_sum, lag0_created_delta_hours_mean

  📌 Remove multicollinear feature
     lag0_created_delta_hours_sum and lag0_created_delta_hours_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_sum, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     lag0_created_delta_hours_sum and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_sum, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_created_delta_hours_sum and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_sum, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_created_delta_hours_sum and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_sum, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     lag0_created_delta_hours_mean and lag0_created_delta_hours_max are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_mean, lag0_created_delta_hours_max

  📌 Remove multicollinear feature
     lag0_created_delta_hours_mean and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_mean, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_created_delta_hours_mean and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_mean, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_created_delta_hours_mean and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_mean, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     lag0_created_delta_hours_max and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_max, created_delta_hours_vs_cohort_mean

  📌 Remove multicollinear feature
     lag0_created_delta_hours_max and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_max, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     lag0_created_delta_hours_max and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: lag0_created_delta_hours_max, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     esent_end and esent_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_end, esent_vs_cohort_mean

  📌 Remove multicollinear feature
     esent_end and esent_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_end, esent_vs_cohort_pct

  📌 Remove multicollinear feature
     esent_end and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_end, esent_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_end and eopenrate_trend_ratio are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, eopenrate_trend_ratio

  📌 Remove multicollinear feature
     eopenrate_end and avgorder_beginning are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, avgorder_beginning

  📌 Remove multicollinear feature
     eopenrate_end and avgorder_end are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, avgorder_end

  📌 Remove multicollinear feature
     eopenrate_end and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_end and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_end and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_end and eclickrate_vs_cohort_mean are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, eclickrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_end and eclickrate_vs_cohort_pct are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_end and eclickrate_cohort_zscore are highly correlated (r=0.94)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_end and avgorder_vs_cohort_mean are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_end and avgorder_vs_cohort_pct are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_end and avgorder_cohort_zscore are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_end, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and avgorder_beginning are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, avgorder_beginning

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and avgorder_end are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, avgorder_end

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and eopenrate_vs_cohort_mean are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and eopenrate_vs_cohort_pct are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and eopenrate_cohort_zscore are highly correlated (r=0.97)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and eclickrate_vs_cohort_mean are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, eclickrate_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and eclickrate_vs_cohort_pct are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and eclickrate_cohort_zscore are highly correlated (r=0.99)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and avgorder_vs_cohort_mean are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and avgorder_vs_cohort_pct are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_trend_ratio and avgorder_cohort_zscore are highly correlated (r=-0.87)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_trend_ratio, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_beginning and avgorder_end are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_beginning, avgorder_end

  📌 Remove multicollinear feature
     avgorder_beginning and eopenrate_vs_cohort_mean are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_beginning, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     avgorder_beginning and eopenrate_vs_cohort_pct are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_beginning, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     avgorder_beginning and eopenrate_cohort_zscore are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_beginning, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_beginning and avgorder_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_beginning, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     avgorder_beginning and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_beginning, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     avgorder_beginning and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_beginning, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_end and eopenrate_vs_cohort_mean are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_end, eopenrate_vs_cohort_mean

  📌 Remove multicollinear feature
     avgorder_end and eopenrate_vs_cohort_pct are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_end, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     avgorder_end and eopenrate_cohort_zscore are highly correlated (r=-0.85)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_end, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_end and avgorder_vs_cohort_mean are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_end, avgorder_vs_cohort_mean

  📌 Remove multicollinear feature
     avgorder_end and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_end, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     avgorder_end and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_end, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     esent_vs_cohort_mean and esent_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_vs_cohort_mean, esent_vs_cohort_pct

  📌 Remove multicollinear feature
     esent_vs_cohort_mean and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_vs_cohort_mean, esent_cohort_zscore

  📌 Remove multicollinear feature
     esent_vs_cohort_pct and esent_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: esent_vs_cohort_pct, esent_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_vs_cohort_mean and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_vs_cohort_mean, eopenrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eopenrate_vs_cohort_mean and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_vs_cohort_mean, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     eopenrate_vs_cohort_pct and eopenrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eopenrate_vs_cohort_pct, eopenrate_cohort_zscore

  📌 Remove multicollinear feature
     eclickrate_vs_cohort_mean and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_vs_cohort_mean, eclickrate_vs_cohort_pct

  📌 Remove multicollinear feature
     eclickrate_vs_cohort_mean and eclickrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_vs_cohort_mean, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     eclickrate_vs_cohort_pct and eclickrate_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: eclickrate_vs_cohort_pct, eclickrate_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_vs_cohort_mean and avgorder_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_vs_cohort_mean, avgorder_vs_cohort_pct

  📌 Remove multicollinear feature
     avgorder_vs_cohort_mean and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_vs_cohort_mean, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     avgorder_vs_cohort_pct and avgorder_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: avgorder_vs_cohort_pct, avgorder_cohort_zscore

  📌 Remove multicollinear feature
     ordfreq_vs_cohort_mean and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_vs_cohort_mean, ordfreq_vs_cohort_pct

  📌 Remove multicollinear feature
     ordfreq_vs_cohort_mean and ordfreq_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_vs_cohort_mean, ordfreq_cohort_zscore

  📌 Remove multicollinear feature
     ordfreq_vs_cohort_pct and ordfreq_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: ordfreq_vs_cohort_pct, ordfreq_cohort_zscore

  📌 Remove multicollinear feature
     created_delta_hours_vs_cohort_mean and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_vs_cohort_mean, created_delta_hours_vs_cohort_pct

  📌 Remove multicollinear feature
     created_delta_hours_vs_cohort_mean and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_vs_cohort_mean, created_delta_hours_cohort_zscore

  📌 Remove multicollinear feature
     created_delta_hours_vs_cohort_pct and created_delta_hours_cohort_zscore are highly correlated (r=-1.00)
     → Action: Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.
     → Features: created_delta_hours_vs_cohort_pct, created_delta_hours_cohort_zscore

  📌 Prioritize strong predictors
     Top predictive features: esent_max_all_time, esent_mean_all_time, esent_vs_cohort_pct
     → Action: Ensure these features are included in your model and check for data quality issues.
     → Features: esent_max_all_time, esent_mean_all_time, esent_vs_cohort_pct

  📌 Stratify by cohort_quarter
     Significant variation in retention rates across cohort_quarter categories (spread: 98.9%)
     → Action: Use stratified sampling by cohort_quarter in train/test split to ensure all segments are represented.
     → Features: cohort_quarter

  📌 Monitor high-risk segments
     Segments with below-average retention: 2008.0Q3.0, 2009.0Q1.0, 2009.0Q2.0
     → Action: Target these segments for intervention campaigns and ensure adequate representation in training data.
     → Features: cohort_quarter, cohort_quarter, cohort_quarter

✅ Persisted 927 multicollinearity recommendations
✅ Persisted 17 strong predictor recommendations
✅ Persisted 10 weak predictor recommendations

# Feature Selection Recommendations
selection_recs = grouped_recs.get(RecommendationCategory.FEATURE_SELECTION, [])

print("=" * 70)
print("FEATURE SELECTION")
print("=" * 70)

# Strong predictors summary
if analysis_summary.strong_predictors:
    print("\n✅ STRONG PREDICTORS (prioritize these):")
    strong_df = pd.DataFrame(analysis_summary.strong_predictors)
    strong_df["effect_size"] = strong_df["effect_size"].apply(lambda x: f"{x:+.3f}")
    strong_df["correlation"] = strong_df["correlation"].apply(lambda x: f"{x:+.3f}")
    strong_df = strong_df.sort_values("effect_size", key=lambda x: x.str.replace("+", "").astype(float).abs(), ascending=False)
    display(strong_df)

    print("\n   💡 These features show strong discrimination between retained/churned customers.")
    print("   → Ensure they're included in your model")
    print("   → Check for data quality issues that could inflate their importance")

# Weak predictors summary
if analysis_summary.weak_predictors:
    print(f"\n⚪ WEAK PREDICTORS (consider dropping): {', '.join(analysis_summary.weak_predictors[:5])}")
    print("   → Low individual predictive power, but may help in combination")

# Multicollinearity summary
if analysis_summary.multicollinear_pairs:
    print("\n⚠️ MULTICOLLINEAR PAIRS (drop one from each pair for linear models):")
    for pair in analysis_summary.multicollinear_pairs:
        print(f"   • {pair['feature1']} ↔ {pair['feature2']}: r = {pair['correlation']:.2f}")
    print("\n   💡 For each pair, keep the feature with:")
    print("      - Stronger business meaning")
    print("      - Higher target correlation")
    print("      - Fewer missing values")

# Display all feature selection recommendations
if selection_recs:
    print("\n" + "-" * 70)
    print("DETAILED RECOMMENDATIONS:")
    for rec in selection_recs:
        priority_icon = "🔴" if rec.priority == "high" else "🟡" if rec.priority == "medium" else "🟢"
        print(f"\n{priority_icon} {rec.title}")
        print(f"   {rec.description}")
        print(f"   → {rec.action}")

======================================================================
FEATURE SELECTION
======================================================================

✅ STRONG PREDICTORS (prioritize these):

   💡 These features show strong discrimination between retained/churned customers.
   → Ensure they're included in your model
   → Check for data quality issues that could inflate their importance

⚪ WEAK PREDICTORS (consider dropping): event_count_all_time, esent_count_all_time, eopenrate_sum_all_time, eopenrate_mean_all_time, eopenrate_max_all_time
   → Low individual predictive power, but may help in combination

⚠️ MULTICOLLINEAR PAIRS (drop one from each pair for linear models):
   • event_count_all_time ↔ esent_count_all_time: r = 1.00
   • event_count_all_time ↔ eopenrate_count_all_time: r = 1.00
   • event_count_all_time ↔ eclickrate_count_all_time: r = 1.00
   • event_count_all_time ↔ avgorder_count_all_time: r = 1.00
   • event_count_all_time ↔ ordfreq_count_all_time: r = 1.00
   • event_count_all_time ↔ paperless_count_all_time: r = 1.00
   • event_count_all_time ↔ refill_count_all_time: r = 1.00
   • event_count_all_time ↔ doorstep_count_all_time: r = 1.00
   • event_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • event_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • event_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • event_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • esent_sum_all_time ↔ esent_mean_all_time: r = 1.00
   • esent_sum_all_time ↔ esent_max_all_time: r = 1.00
   • esent_sum_all_time ↔ lag0_esent_sum: r = 1.00
   • esent_sum_all_time ↔ lag0_esent_mean: r = 1.00
   • esent_sum_all_time ↔ lag0_esent_max: r = 1.00
   • esent_sum_all_time ↔ esent_beginning: r = 0.90
   • esent_sum_all_time ↔ esent_end: r = 0.94
   • esent_sum_all_time ↔ esent_vs_cohort_mean: r = 1.00
   • esent_sum_all_time ↔ esent_vs_cohort_pct: r = 1.00
   • esent_sum_all_time ↔ esent_cohort_zscore: r = 1.00
   • esent_mean_all_time ↔ esent_max_all_time: r = 1.00
   • esent_mean_all_time ↔ lag0_esent_sum: r = 1.00
   • esent_mean_all_time ↔ lag0_esent_mean: r = 1.00
   • esent_mean_all_time ↔ lag0_esent_max: r = 1.00
   • esent_mean_all_time ↔ esent_beginning: r = 0.90
   • esent_mean_all_time ↔ esent_end: r = 0.94
   • esent_mean_all_time ↔ esent_vs_cohort_mean: r = 1.00
   • esent_mean_all_time ↔ esent_vs_cohort_pct: r = 1.00
   • esent_mean_all_time ↔ esent_cohort_zscore: r = 1.00
   • esent_max_all_time ↔ lag0_esent_sum: r = 1.00
   • esent_max_all_time ↔ lag0_esent_mean: r = 1.00
   • esent_max_all_time ↔ lag0_esent_max: r = 1.00
   • esent_max_all_time ↔ esent_beginning: r = 0.98
   • esent_max_all_time ↔ esent_end: r = 0.82
   • esent_max_all_time ↔ esent_vs_cohort_mean: r = 1.00
   • esent_max_all_time ↔ esent_vs_cohort_pct: r = 1.00
   • esent_max_all_time ↔ esent_cohort_zscore: r = 1.00
   • esent_count_all_time ↔ eopenrate_count_all_time: r = 1.00
   • esent_count_all_time ↔ eclickrate_count_all_time: r = 1.00
   • esent_count_all_time ↔ avgorder_count_all_time: r = 1.00
   • esent_count_all_time ↔ ordfreq_count_all_time: r = 1.00
   • esent_count_all_time ↔ paperless_count_all_time: r = 1.00
   • esent_count_all_time ↔ refill_count_all_time: r = 1.00
   • esent_count_all_time ↔ doorstep_count_all_time: r = 1.00
   • esent_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • esent_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • esent_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • esent_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • eopenrate_sum_all_time ↔ eopenrate_mean_all_time: r = 1.00
   • eopenrate_sum_all_time ↔ eopenrate_max_all_time: r = 1.00
   • eopenrate_sum_all_time ↔ lag0_eopenrate_sum: r = 1.00
   • eopenrate_sum_all_time ↔ lag0_eopenrate_mean: r = 1.00
   • eopenrate_sum_all_time ↔ lag0_eopenrate_max: r = 1.00
   • eopenrate_sum_all_time ↔ eopenrate_beginning: r = 0.89
   • eopenrate_sum_all_time ↔ eopenrate_vs_cohort_mean: r = 1.00
   • eopenrate_sum_all_time ↔ eopenrate_vs_cohort_pct: r = 1.00
   • eopenrate_sum_all_time ↔ eopenrate_cohort_zscore: r = 1.00
   • eopenrate_mean_all_time ↔ eopenrate_max_all_time: r = 1.00
   • eopenrate_mean_all_time ↔ lag0_eopenrate_sum: r = 1.00
   • eopenrate_mean_all_time ↔ lag0_eopenrate_mean: r = 1.00
   • eopenrate_mean_all_time ↔ lag0_eopenrate_max: r = 1.00
   • eopenrate_mean_all_time ↔ eopenrate_beginning: r = 0.89
   • eopenrate_mean_all_time ↔ eopenrate_vs_cohort_mean: r = 1.00
   • eopenrate_mean_all_time ↔ eopenrate_vs_cohort_pct: r = 1.00
   • eopenrate_mean_all_time ↔ eopenrate_cohort_zscore: r = 1.00
   • eopenrate_max_all_time ↔ lag0_eopenrate_sum: r = 1.00
   • eopenrate_max_all_time ↔ lag0_eopenrate_mean: r = 1.00
   • eopenrate_max_all_time ↔ lag0_eopenrate_max: r = 1.00
   • eopenrate_max_all_time ↔ eopenrate_beginning: r = 0.96
   • eopenrate_max_all_time ↔ eopenrate_vs_cohort_mean: r = 1.00
   • eopenrate_max_all_time ↔ eopenrate_vs_cohort_pct: r = 1.00
   • eopenrate_max_all_time ↔ eopenrate_cohort_zscore: r = 1.00
   • eopenrate_count_all_time ↔ eclickrate_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ avgorder_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ ordfreq_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ paperless_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ refill_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ doorstep_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • eopenrate_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • eclickrate_sum_all_time ↔ eclickrate_mean_all_time: r = 1.00
   • eclickrate_sum_all_time ↔ eclickrate_max_all_time: r = 1.00
   • eclickrate_sum_all_time ↔ lag0_eclickrate_sum: r = 1.00
   • eclickrate_sum_all_time ↔ lag0_eclickrate_mean: r = 1.00
   • eclickrate_sum_all_time ↔ lag0_eclickrate_max: r = 1.00
   • eclickrate_sum_all_time ↔ eopenrate_end: r = 0.96
   • eclickrate_sum_all_time ↔ eopenrate_trend_ratio: r = 0.90
   • eclickrate_sum_all_time ↔ avgorder_beginning: r = -0.93
   • eclickrate_sum_all_time ↔ avgorder_end: r = -0.93
   • eclickrate_sum_all_time ↔ eclickrate_vs_cohort_mean: r = 1.00
   • eclickrate_sum_all_time ↔ eclickrate_vs_cohort_pct: r = 1.00
   • eclickrate_sum_all_time ↔ eclickrate_cohort_zscore: r = 1.00
   • eclickrate_mean_all_time ↔ eclickrate_max_all_time: r = 1.00
   • eclickrate_mean_all_time ↔ lag0_eclickrate_sum: r = 1.00
   • eclickrate_mean_all_time ↔ lag0_eclickrate_mean: r = 1.00
   • eclickrate_mean_all_time ↔ lag0_eclickrate_max: r = 1.00
   • eclickrate_mean_all_time ↔ eopenrate_end: r = 0.96
   • eclickrate_mean_all_time ↔ eopenrate_trend_ratio: r = 0.90
   • eclickrate_mean_all_time ↔ avgorder_beginning: r = -0.93
   • eclickrate_mean_all_time ↔ avgorder_end: r = -0.93
   • eclickrate_mean_all_time ↔ eclickrate_vs_cohort_mean: r = 1.00
   • eclickrate_mean_all_time ↔ eclickrate_vs_cohort_pct: r = 1.00
   • eclickrate_mean_all_time ↔ eclickrate_cohort_zscore: r = 1.00
   • eclickrate_max_all_time ↔ lag0_eclickrate_sum: r = 1.00
   • eclickrate_max_all_time ↔ lag0_eclickrate_mean: r = 1.00
   • eclickrate_max_all_time ↔ lag0_eclickrate_max: r = 1.00
   • eclickrate_max_all_time ↔ eopenrate_end: r = 0.96
   • eclickrate_max_all_time ↔ eopenrate_trend_ratio: r = 0.90
   • eclickrate_max_all_time ↔ avgorder_beginning: r = -0.93
   • eclickrate_max_all_time ↔ avgorder_end: r = -0.93
   • eclickrate_max_all_time ↔ eclickrate_vs_cohort_mean: r = 1.00
   • eclickrate_max_all_time ↔ eclickrate_vs_cohort_pct: r = 1.00
   • eclickrate_max_all_time ↔ eclickrate_cohort_zscore: r = 1.00
   • eclickrate_count_all_time ↔ avgorder_count_all_time: r = 1.00
   • eclickrate_count_all_time ↔ ordfreq_count_all_time: r = 1.00
   • eclickrate_count_all_time ↔ paperless_count_all_time: r = 1.00
   • eclickrate_count_all_time ↔ refill_count_all_time: r = 1.00
   • eclickrate_count_all_time ↔ doorstep_count_all_time: r = 1.00
   • eclickrate_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • eclickrate_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • eclickrate_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • eclickrate_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • avgorder_sum_all_time ↔ avgorder_mean_all_time: r = 1.00
   • avgorder_sum_all_time ↔ avgorder_max_all_time: r = 1.00
   • avgorder_sum_all_time ↔ lag0_avgorder_sum: r = 1.00
   • avgorder_sum_all_time ↔ lag0_avgorder_mean: r = 1.00
   • avgorder_sum_all_time ↔ lag0_avgorder_max: r = 1.00
   • avgorder_sum_all_time ↔ esent_trend_ratio: r = -0.76
   • avgorder_sum_all_time ↔ eopenrate_end: r = -0.85
   • avgorder_sum_all_time ↔ eopenrate_trend_ratio: r = -0.87
   • avgorder_sum_all_time ↔ avgorder_beginning: r = 1.00
   • avgorder_sum_all_time ↔ avgorder_end: r = 1.00
   • avgorder_sum_all_time ↔ avgorder_vs_cohort_mean: r = 1.00
   • avgorder_sum_all_time ↔ avgorder_vs_cohort_pct: r = 1.00
   • avgorder_sum_all_time ↔ avgorder_cohort_zscore: r = 1.00
   • avgorder_mean_all_time ↔ avgorder_max_all_time: r = 1.00
   • avgorder_mean_all_time ↔ lag0_avgorder_sum: r = 1.00
   • avgorder_mean_all_time ↔ lag0_avgorder_mean: r = 1.00
   • avgorder_mean_all_time ↔ lag0_avgorder_max: r = 1.00
   • avgorder_mean_all_time ↔ esent_trend_ratio: r = -0.76
   • avgorder_mean_all_time ↔ eopenrate_end: r = -0.85
   • avgorder_mean_all_time ↔ eopenrate_trend_ratio: r = -0.87
   • avgorder_mean_all_time ↔ avgorder_beginning: r = 1.00
   • avgorder_mean_all_time ↔ avgorder_end: r = 1.00
   • avgorder_mean_all_time ↔ avgorder_vs_cohort_mean: r = 1.00
   • avgorder_mean_all_time ↔ avgorder_vs_cohort_pct: r = 1.00
   • avgorder_mean_all_time ↔ avgorder_cohort_zscore: r = 1.00
   • avgorder_max_all_time ↔ lag0_avgorder_sum: r = 1.00
   • avgorder_max_all_time ↔ lag0_avgorder_mean: r = 1.00
   • avgorder_max_all_time ↔ lag0_avgorder_max: r = 1.00
   • avgorder_max_all_time ↔ esent_trend_ratio: r = -0.76
   • avgorder_max_all_time ↔ eopenrate_end: r = -0.85
   • avgorder_max_all_time ↔ eopenrate_trend_ratio: r = -0.87
   • avgorder_max_all_time ↔ avgorder_beginning: r = 1.00
   • avgorder_max_all_time ↔ avgorder_end: r = 1.00
   • avgorder_max_all_time ↔ avgorder_vs_cohort_mean: r = 1.00
   • avgorder_max_all_time ↔ avgorder_vs_cohort_pct: r = 1.00
   • avgorder_max_all_time ↔ avgorder_cohort_zscore: r = 1.00
   • avgorder_count_all_time ↔ ordfreq_count_all_time: r = 1.00
   • avgorder_count_all_time ↔ paperless_count_all_time: r = 1.00
   • avgorder_count_all_time ↔ refill_count_all_time: r = 1.00
   • avgorder_count_all_time ↔ doorstep_count_all_time: r = 1.00
   • avgorder_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • avgorder_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • avgorder_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • avgorder_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • ordfreq_sum_all_time ↔ ordfreq_mean_all_time: r = 1.00
   • ordfreq_sum_all_time ↔ ordfreq_max_all_time: r = 1.00
   • ordfreq_sum_all_time ↔ lag0_ordfreq_sum: r = 1.00
   • ordfreq_sum_all_time ↔ lag0_ordfreq_mean: r = 1.00
   • ordfreq_sum_all_time ↔ lag0_ordfreq_max: r = 1.00
   • ordfreq_sum_all_time ↔ ordfreq_vs_cohort_mean: r = 1.00
   • ordfreq_sum_all_time ↔ ordfreq_vs_cohort_pct: r = 1.00
   • ordfreq_sum_all_time ↔ ordfreq_cohort_zscore: r = 1.00
   • ordfreq_mean_all_time ↔ ordfreq_max_all_time: r = 1.00
   • ordfreq_mean_all_time ↔ lag0_ordfreq_sum: r = 1.00
   • ordfreq_mean_all_time ↔ lag0_ordfreq_mean: r = 1.00
   • ordfreq_mean_all_time ↔ lag0_ordfreq_max: r = 1.00
   • ordfreq_mean_all_time ↔ ordfreq_vs_cohort_mean: r = 1.00
   • ordfreq_mean_all_time ↔ ordfreq_vs_cohort_pct: r = 1.00
   • ordfreq_mean_all_time ↔ ordfreq_cohort_zscore: r = 1.00
   • ordfreq_max_all_time ↔ lag0_ordfreq_sum: r = 1.00
   • ordfreq_max_all_time ↔ lag0_ordfreq_mean: r = 1.00
   • ordfreq_max_all_time ↔ lag0_ordfreq_max: r = 1.00
   • ordfreq_max_all_time ↔ ordfreq_vs_cohort_mean: r = 1.00
   • ordfreq_max_all_time ↔ ordfreq_vs_cohort_pct: r = 1.00
   • ordfreq_max_all_time ↔ ordfreq_cohort_zscore: r = 1.00
   • ordfreq_count_all_time ↔ paperless_count_all_time: r = 1.00
   • ordfreq_count_all_time ↔ refill_count_all_time: r = 1.00
   • ordfreq_count_all_time ↔ doorstep_count_all_time: r = 1.00
   • ordfreq_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • ordfreq_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • ordfreq_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • ordfreq_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • paperless_sum_all_time ↔ esent_trend_ratio: r = 0.87
   • paperless_sum_all_time ↔ avgorder_beginning: r = -0.92
   • paperless_sum_all_time ↔ avgorder_end: r = -0.92
   • paperless_count_all_time ↔ refill_count_all_time: r = 1.00
   • paperless_count_all_time ↔ doorstep_count_all_time: r = 1.00
   • paperless_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • paperless_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • paperless_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • paperless_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • refill_count_all_time ↔ doorstep_count_all_time: r = 1.00
   • refill_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • refill_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • refill_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • refill_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • doorstep_count_all_time ↔ firstorder_is_weekend_count_all_time: r = 1.00
   • doorstep_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • doorstep_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • doorstep_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • created_delta_hours_sum_all_time ↔ created_delta_hours_mean_all_time: r = 1.00
   • created_delta_hours_sum_all_time ↔ created_delta_hours_max_all_time: r = 1.00
   • created_delta_hours_sum_all_time ↔ days_since_created_sum_all_time: r = -1.00
   • created_delta_hours_sum_all_time ↔ days_since_created_mean_all_time: r = -1.00
   • created_delta_hours_sum_all_time ↔ days_since_created_max_all_time: r = -1.00
   • created_delta_hours_sum_all_time ↔ days_until_created_sum_all_time: r = 1.00
   • created_delta_hours_sum_all_time ↔ days_until_created_mean_all_time: r = 1.00
   • created_delta_hours_sum_all_time ↔ days_until_created_max_all_time: r = 1.00
   • created_delta_hours_sum_all_time ↔ log1p_days_since_created_sum_all_time: r = -0.72
   • created_delta_hours_sum_all_time ↔ log1p_days_since_created_mean_all_time: r = -0.71
   • created_delta_hours_sum_all_time ↔ log1p_days_since_created_max_all_time: r = -0.71
   • created_delta_hours_sum_all_time ↔ days_since_firstorder_sum_all_time: r = -0.89
   • created_delta_hours_sum_all_time ↔ days_since_firstorder_mean_all_time: r = -0.89
   • created_delta_hours_sum_all_time ↔ days_since_firstorder_max_all_time: r = -0.89
   • created_delta_hours_sum_all_time ↔ days_until_firstorder_sum_all_time: r = 0.89
   • created_delta_hours_sum_all_time ↔ days_until_firstorder_mean_all_time: r = 0.89
   • created_delta_hours_sum_all_time ↔ days_until_firstorder_max_all_time: r = 0.89
   • created_delta_hours_sum_all_time ↔ lastorder_delta_hours_sum_all_time: r = -0.95
   • created_delta_hours_sum_all_time ↔ lastorder_delta_hours_mean_all_time: r = -0.95
   • created_delta_hours_sum_all_time ↔ lastorder_delta_hours_max_all_time: r = -0.95
   • created_delta_hours_sum_all_time ↔ lag0_created_delta_hours_sum: r = 1.00
   • created_delta_hours_sum_all_time ↔ lag0_created_delta_hours_mean: r = 1.00
   • created_delta_hours_sum_all_time ↔ lag0_created_delta_hours_max: r = 1.00
   • created_delta_hours_sum_all_time ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • created_delta_hours_sum_all_time ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • created_delta_hours_sum_all_time ↔ created_delta_hours_cohort_zscore: r = 1.00
   • created_delta_hours_mean_all_time ↔ created_delta_hours_max_all_time: r = 1.00
   • created_delta_hours_mean_all_time ↔ days_since_created_sum_all_time: r = -1.00
   • created_delta_hours_mean_all_time ↔ days_since_created_mean_all_time: r = -1.00
   • created_delta_hours_mean_all_time ↔ days_since_created_max_all_time: r = -1.00
   • created_delta_hours_mean_all_time ↔ days_until_created_sum_all_time: r = 1.00
   • created_delta_hours_mean_all_time ↔ days_until_created_mean_all_time: r = 1.00
   • created_delta_hours_mean_all_time ↔ days_until_created_max_all_time: r = 1.00
   • created_delta_hours_mean_all_time ↔ log1p_days_since_created_sum_all_time: r = -0.71
   • created_delta_hours_mean_all_time ↔ log1p_days_since_created_mean_all_time: r = -0.71
   • created_delta_hours_mean_all_time ↔ log1p_days_since_created_max_all_time: r = -0.71
   • created_delta_hours_mean_all_time ↔ days_since_firstorder_sum_all_time: r = -0.89
   • created_delta_hours_mean_all_time ↔ days_since_firstorder_mean_all_time: r = -0.89
   • created_delta_hours_mean_all_time ↔ days_since_firstorder_max_all_time: r = -0.89
   • created_delta_hours_mean_all_time ↔ days_until_firstorder_sum_all_time: r = 0.89
   • created_delta_hours_mean_all_time ↔ days_until_firstorder_mean_all_time: r = 0.89
   • created_delta_hours_mean_all_time ↔ days_until_firstorder_max_all_time: r = 0.89
   • created_delta_hours_mean_all_time ↔ lastorder_delta_hours_sum_all_time: r = -1.00
   • created_delta_hours_mean_all_time ↔ lastorder_delta_hours_mean_all_time: r = -1.00
   • created_delta_hours_mean_all_time ↔ lastorder_delta_hours_max_all_time: r = -1.00
   • created_delta_hours_mean_all_time ↔ lag0_created_delta_hours_sum: r = 1.00
   • created_delta_hours_mean_all_time ↔ lag0_created_delta_hours_mean: r = 1.00
   • created_delta_hours_mean_all_time ↔ lag0_created_delta_hours_max: r = 1.00
   • created_delta_hours_mean_all_time ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • created_delta_hours_mean_all_time ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • created_delta_hours_mean_all_time ↔ created_delta_hours_cohort_zscore: r = 1.00
   • created_delta_hours_max_all_time ↔ days_since_created_sum_all_time: r = -1.00
   • created_delta_hours_max_all_time ↔ days_since_created_mean_all_time: r = -1.00
   • created_delta_hours_max_all_time ↔ days_since_created_max_all_time: r = -1.00
   • created_delta_hours_max_all_time ↔ days_until_created_sum_all_time: r = 1.00
   • created_delta_hours_max_all_time ↔ days_until_created_mean_all_time: r = 1.00
   • created_delta_hours_max_all_time ↔ days_until_created_max_all_time: r = 1.00
   • created_delta_hours_max_all_time ↔ log1p_days_since_created_sum_all_time: r = -0.71
   • created_delta_hours_max_all_time ↔ log1p_days_since_created_mean_all_time: r = -0.71
   • created_delta_hours_max_all_time ↔ log1p_days_since_created_max_all_time: r = -0.71
   • created_delta_hours_max_all_time ↔ days_since_firstorder_sum_all_time: r = -0.89
   • created_delta_hours_max_all_time ↔ days_since_firstorder_mean_all_time: r = -0.89
   • created_delta_hours_max_all_time ↔ days_since_firstorder_max_all_time: r = -0.89
   • created_delta_hours_max_all_time ↔ days_until_firstorder_sum_all_time: r = 0.89
   • created_delta_hours_max_all_time ↔ days_until_firstorder_mean_all_time: r = 0.89
   • created_delta_hours_max_all_time ↔ days_until_firstorder_max_all_time: r = 0.89
   • created_delta_hours_max_all_time ↔ lastorder_delta_hours_sum_all_time: r = -1.00
   • created_delta_hours_max_all_time ↔ lastorder_delta_hours_mean_all_time: r = -1.00
   • created_delta_hours_max_all_time ↔ lastorder_delta_hours_max_all_time: r = -1.00
   • created_delta_hours_max_all_time ↔ lag0_created_delta_hours_sum: r = 1.00
   • created_delta_hours_max_all_time ↔ lag0_created_delta_hours_mean: r = 1.00
   • created_delta_hours_max_all_time ↔ lag0_created_delta_hours_max: r = 1.00
   • created_delta_hours_max_all_time ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • created_delta_hours_max_all_time ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • created_delta_hours_max_all_time ↔ created_delta_hours_cohort_zscore: r = 1.00
   • created_delta_hours_count_all_time ↔ created_hour_count_all_time: r = 0.97
   • created_delta_hours_count_all_time ↔ created_dow_count_all_time: r = 0.97
   • created_delta_hours_count_all_time ↔ created_is_weekend_count_all_time: r = 0.97
   • created_delta_hours_count_all_time ↔ days_since_created_count_all_time: r = 1.00
   • created_delta_hours_count_all_time ↔ days_until_created_count_all_time: r = 1.00
   • created_delta_hours_count_all_time ↔ log1p_days_since_created_count_all_time: r = 1.00
   • created_delta_hours_count_all_time ↔ is_future_created_count_all_time: r = 0.97
   • created_delta_hours_count_all_time ↔ eopenrate_end: r = 0.94
   • created_delta_hours_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • created_delta_hours_count_all_time ↔ avgorder_beginning: r = -0.79
   • created_delta_hours_count_all_time ↔ avgorder_end: r = -0.79
   • created_hour_count_all_time ↔ created_dow_count_all_time: r = 1.00
   • created_hour_count_all_time ↔ created_is_weekend_count_all_time: r = 1.00
   • created_hour_count_all_time ↔ days_since_created_count_all_time: r = 0.97
   • created_hour_count_all_time ↔ days_until_created_count_all_time: r = 0.97
   • created_hour_count_all_time ↔ log1p_days_since_created_count_all_time: r = 0.97
   • created_hour_count_all_time ↔ is_future_created_count_all_time: r = 1.00
   • created_hour_count_all_time ↔ eopenrate_end: r = 0.94
   • created_hour_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • created_hour_count_all_time ↔ avgorder_beginning: r = -0.79
   • created_hour_count_all_time ↔ avgorder_end: r = -0.79
   • created_dow_sum_all_time ↔ created_dow_mean_all_time: r = 1.00
   • created_dow_sum_all_time ↔ created_dow_max_all_time: r = 1.00
   • created_dow_sum_all_time ↔ dow_sin: r = -0.71
   • created_dow_sum_all_time ↔ esent_beginning: r = -0.76
   • created_dow_mean_all_time ↔ created_dow_max_all_time: r = 1.00
   • created_dow_mean_all_time ↔ dow_sin: r = -0.71
   • created_dow_max_all_time ↔ dow_sin: r = -0.71
   • created_dow_count_all_time ↔ created_is_weekend_count_all_time: r = 1.00
   • created_dow_count_all_time ↔ days_since_created_count_all_time: r = 0.97
   • created_dow_count_all_time ↔ days_until_created_count_all_time: r = 0.97
   • created_dow_count_all_time ↔ log1p_days_since_created_count_all_time: r = 0.97
   • created_dow_count_all_time ↔ is_future_created_count_all_time: r = 1.00
   • created_dow_count_all_time ↔ eopenrate_end: r = 0.94
   • created_dow_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • created_dow_count_all_time ↔ avgorder_beginning: r = -0.79
   • created_dow_count_all_time ↔ avgorder_end: r = -0.79
   • created_is_weekend_count_all_time ↔ days_since_created_count_all_time: r = 0.97
   • created_is_weekend_count_all_time ↔ days_until_created_count_all_time: r = 0.97
   • created_is_weekend_count_all_time ↔ log1p_days_since_created_count_all_time: r = 0.97
   • created_is_weekend_count_all_time ↔ is_future_created_count_all_time: r = 1.00
   • created_is_weekend_count_all_time ↔ eopenrate_end: r = 0.94
   • created_is_weekend_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • created_is_weekend_count_all_time ↔ avgorder_beginning: r = -0.79
   • created_is_weekend_count_all_time ↔ avgorder_end: r = -0.79
   • firstorder_delta_hours_sum_all_time ↔ firstorder_delta_hours_mean_all_time: r = 1.00
   • firstorder_delta_hours_sum_all_time ↔ firstorder_delta_hours_max_all_time: r = 1.00
   • firstorder_delta_hours_sum_all_time ↔ eopenrate_beginning: r = 0.93
   • firstorder_delta_hours_mean_all_time ↔ firstorder_delta_hours_max_all_time: r = 1.00
   • firstorder_delta_hours_mean_all_time ↔ eopenrate_beginning: r = 0.93
   • firstorder_delta_hours_max_all_time ↔ eopenrate_beginning: r = 0.93
   • firstorder_delta_hours_count_all_time ↔ firstorder_hour_count_all_time: r = 1.00
   • firstorder_delta_hours_count_all_time ↔ firstorder_dow_count_all_time: r = 1.00
   • firstorder_delta_hours_count_all_time ↔ lastorder_delta_hours_count_all_time: r = 0.76
   • firstorder_delta_hours_count_all_time ↔ lastorder_hour_count_all_time: r = 0.76
   • firstorder_delta_hours_count_all_time ↔ lastorder_dow_count_all_time: r = 0.76
   • firstorder_hour_count_all_time ↔ firstorder_dow_count_all_time: r = 1.00
   • firstorder_hour_count_all_time ↔ lastorder_delta_hours_count_all_time: r = 0.76
   • firstorder_hour_count_all_time ↔ lastorder_hour_count_all_time: r = 0.76
   • firstorder_hour_count_all_time ↔ lastorder_dow_count_all_time: r = 0.76
   • firstorder_dow_sum_all_time ↔ firstorder_dow_mean_all_time: r = 1.00
   • firstorder_dow_sum_all_time ↔ firstorder_dow_max_all_time: r = 1.00
   • firstorder_dow_sum_all_time ↔ firstorder_is_weekend_mean_all_time: r = 0.79
   • firstorder_dow_sum_all_time ↔ eclickrate_beginning: r = 0.81
   • firstorder_dow_mean_all_time ↔ firstorder_dow_max_all_time: r = 1.00
   • firstorder_dow_mean_all_time ↔ firstorder_is_weekend_mean_all_time: r = 0.79
   • firstorder_dow_mean_all_time ↔ eclickrate_beginning: r = 0.81
   • firstorder_dow_max_all_time ↔ firstorder_is_weekend_mean_all_time: r = 0.79
   • firstorder_dow_max_all_time ↔ eclickrate_beginning: r = 0.81
   • firstorder_dow_count_all_time ↔ lastorder_delta_hours_count_all_time: r = 0.76
   • firstorder_dow_count_all_time ↔ lastorder_hour_count_all_time: r = 0.76
   • firstorder_dow_count_all_time ↔ lastorder_dow_count_all_time: r = 0.76
   • firstorder_is_weekend_count_all_time ↔ is_missing_created_count_all_time: r = 1.00
   • firstorder_is_weekend_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • firstorder_is_weekend_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • days_since_created_sum_all_time ↔ days_since_created_mean_all_time: r = 1.00
   • days_since_created_sum_all_time ↔ days_since_created_max_all_time: r = 1.00
   • days_since_created_sum_all_time ↔ days_until_created_sum_all_time: r = -1.00
   • days_since_created_sum_all_time ↔ days_until_created_mean_all_time: r = -1.00
   • days_since_created_sum_all_time ↔ days_until_created_max_all_time: r = -1.00
   • days_since_created_sum_all_time ↔ log1p_days_since_created_sum_all_time: r = 0.72
   • days_since_created_sum_all_time ↔ log1p_days_since_created_mean_all_time: r = 0.71
   • days_since_created_sum_all_time ↔ log1p_days_since_created_max_all_time: r = 0.71
   • days_since_created_sum_all_time ↔ days_since_firstorder_sum_all_time: r = 0.89
   • days_since_created_sum_all_time ↔ days_since_firstorder_mean_all_time: r = 0.89
   • days_since_created_sum_all_time ↔ days_since_firstorder_max_all_time: r = 0.89
   • days_since_created_sum_all_time ↔ days_until_firstorder_sum_all_time: r = -0.89
   • days_since_created_sum_all_time ↔ days_until_firstorder_mean_all_time: r = -0.89
   • days_since_created_sum_all_time ↔ days_until_firstorder_max_all_time: r = -0.89
   • days_since_created_sum_all_time ↔ lastorder_delta_hours_sum_all_time: r = 0.95
   • days_since_created_sum_all_time ↔ lastorder_delta_hours_mean_all_time: r = 0.95
   • days_since_created_sum_all_time ↔ lastorder_delta_hours_max_all_time: r = 0.95
   • days_since_created_sum_all_time ↔ lag0_created_delta_hours_sum: r = -1.00
   • days_since_created_sum_all_time ↔ lag0_created_delta_hours_mean: r = -1.00
   • days_since_created_sum_all_time ↔ lag0_created_delta_hours_max: r = -1.00
   • days_since_created_sum_all_time ↔ created_delta_hours_vs_cohort_mean: r = -1.00
   • days_since_created_sum_all_time ↔ created_delta_hours_vs_cohort_pct: r = 1.00
   • days_since_created_sum_all_time ↔ created_delta_hours_cohort_zscore: r = -1.00
   • days_since_created_mean_all_time ↔ days_since_created_max_all_time: r = 1.00
   • days_since_created_mean_all_time ↔ days_until_created_sum_all_time: r = -1.00
   • days_since_created_mean_all_time ↔ days_until_created_mean_all_time: r = -1.00
   • days_since_created_mean_all_time ↔ days_until_created_max_all_time: r = -1.00
   • days_since_created_mean_all_time ↔ log1p_days_since_created_sum_all_time: r = 0.71
   • days_since_created_mean_all_time ↔ log1p_days_since_created_mean_all_time: r = 0.71
   • days_since_created_mean_all_time ↔ log1p_days_since_created_max_all_time: r = 0.71
   • days_since_created_mean_all_time ↔ days_since_firstorder_sum_all_time: r = 0.89
   • days_since_created_mean_all_time ↔ days_since_firstorder_mean_all_time: r = 0.89
   • days_since_created_mean_all_time ↔ days_since_firstorder_max_all_time: r = 0.89
   • days_since_created_mean_all_time ↔ days_until_firstorder_sum_all_time: r = -0.89
   • days_since_created_mean_all_time ↔ days_until_firstorder_mean_all_time: r = -0.89
   • days_since_created_mean_all_time ↔ days_until_firstorder_max_all_time: r = -0.89
   • days_since_created_mean_all_time ↔ lastorder_delta_hours_sum_all_time: r = 1.00
   • days_since_created_mean_all_time ↔ lastorder_delta_hours_mean_all_time: r = 1.00
   • days_since_created_mean_all_time ↔ lastorder_delta_hours_max_all_time: r = 1.00
   • days_since_created_mean_all_time ↔ lag0_created_delta_hours_sum: r = -1.00
   • days_since_created_mean_all_time ↔ lag0_created_delta_hours_mean: r = -1.00
   • days_since_created_mean_all_time ↔ lag0_created_delta_hours_max: r = -1.00
   • days_since_created_mean_all_time ↔ created_delta_hours_vs_cohort_mean: r = -1.00
   • days_since_created_mean_all_time ↔ created_delta_hours_vs_cohort_pct: r = 1.00
   • days_since_created_mean_all_time ↔ created_delta_hours_cohort_zscore: r = -1.00
   • days_since_created_max_all_time ↔ days_until_created_sum_all_time: r = -1.00
   • days_since_created_max_all_time ↔ days_until_created_mean_all_time: r = -1.00
   • days_since_created_max_all_time ↔ days_until_created_max_all_time: r = -1.00
   • days_since_created_max_all_time ↔ log1p_days_since_created_sum_all_time: r = 0.71
   • days_since_created_max_all_time ↔ log1p_days_since_created_mean_all_time: r = 0.71
   • days_since_created_max_all_time ↔ log1p_days_since_created_max_all_time: r = 0.71
   • days_since_created_max_all_time ↔ days_since_firstorder_sum_all_time: r = 0.89
   • days_since_created_max_all_time ↔ days_since_firstorder_mean_all_time: r = 0.89
   • days_since_created_max_all_time ↔ days_since_firstorder_max_all_time: r = 0.89
   • days_since_created_max_all_time ↔ days_until_firstorder_sum_all_time: r = -0.89
   • days_since_created_max_all_time ↔ days_until_firstorder_mean_all_time: r = -0.89
   • days_since_created_max_all_time ↔ days_until_firstorder_max_all_time: r = -0.89
   • days_since_created_max_all_time ↔ lastorder_delta_hours_sum_all_time: r = 1.00
   • days_since_created_max_all_time ↔ lastorder_delta_hours_mean_all_time: r = 1.00
   • days_since_created_max_all_time ↔ lastorder_delta_hours_max_all_time: r = 1.00
   • days_since_created_max_all_time ↔ lag0_created_delta_hours_sum: r = -1.00
   • days_since_created_max_all_time ↔ lag0_created_delta_hours_mean: r = -1.00
   • days_since_created_max_all_time ↔ lag0_created_delta_hours_max: r = -1.00
   • days_since_created_max_all_time ↔ created_delta_hours_vs_cohort_mean: r = -1.00
   • days_since_created_max_all_time ↔ created_delta_hours_vs_cohort_pct: r = 1.00
   • days_since_created_max_all_time ↔ created_delta_hours_cohort_zscore: r = -1.00
   • days_since_created_count_all_time ↔ days_until_created_count_all_time: r = 1.00
   • days_since_created_count_all_time ↔ log1p_days_since_created_count_all_time: r = 1.00
   • days_since_created_count_all_time ↔ is_future_created_count_all_time: r = 0.97
   • days_since_created_count_all_time ↔ eopenrate_end: r = 0.94
   • days_since_created_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • days_since_created_count_all_time ↔ avgorder_beginning: r = -0.79
   • days_since_created_count_all_time ↔ avgorder_end: r = -0.79
   • days_until_created_sum_all_time ↔ days_until_created_mean_all_time: r = 1.00
   • days_until_created_sum_all_time ↔ days_until_created_max_all_time: r = 1.00
   • days_until_created_sum_all_time ↔ log1p_days_since_created_sum_all_time: r = -0.72
   • days_until_created_sum_all_time ↔ log1p_days_since_created_mean_all_time: r = -0.71
   • days_until_created_sum_all_time ↔ log1p_days_since_created_max_all_time: r = -0.71
   • days_until_created_sum_all_time ↔ days_since_firstorder_sum_all_time: r = -0.89
   • days_until_created_sum_all_time ↔ days_since_firstorder_mean_all_time: r = -0.89
   • days_until_created_sum_all_time ↔ days_since_firstorder_max_all_time: r = -0.89
   • days_until_created_sum_all_time ↔ days_until_firstorder_sum_all_time: r = 0.89
   • days_until_created_sum_all_time ↔ days_until_firstorder_mean_all_time: r = 0.89
   • days_until_created_sum_all_time ↔ days_until_firstorder_max_all_time: r = 0.89
   • days_until_created_sum_all_time ↔ lastorder_delta_hours_sum_all_time: r = -0.95
   • days_until_created_sum_all_time ↔ lastorder_delta_hours_mean_all_time: r = -0.95
   • days_until_created_sum_all_time ↔ lastorder_delta_hours_max_all_time: r = -0.95
   • days_until_created_sum_all_time ↔ lag0_created_delta_hours_sum: r = 1.00
   • days_until_created_sum_all_time ↔ lag0_created_delta_hours_mean: r = 1.00
   • days_until_created_sum_all_time ↔ lag0_created_delta_hours_max: r = 1.00
   • days_until_created_sum_all_time ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • days_until_created_sum_all_time ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • days_until_created_sum_all_time ↔ created_delta_hours_cohort_zscore: r = 1.00
   • days_until_created_mean_all_time ↔ days_until_created_max_all_time: r = 1.00
   • days_until_created_mean_all_time ↔ log1p_days_since_created_sum_all_time: r = -0.71
   • days_until_created_mean_all_time ↔ log1p_days_since_created_mean_all_time: r = -0.71
   • days_until_created_mean_all_time ↔ log1p_days_since_created_max_all_time: r = -0.71
   • days_until_created_mean_all_time ↔ days_since_firstorder_sum_all_time: r = -0.89
   • days_until_created_mean_all_time ↔ days_since_firstorder_mean_all_time: r = -0.89
   • days_until_created_mean_all_time ↔ days_since_firstorder_max_all_time: r = -0.89
   • days_until_created_mean_all_time ↔ days_until_firstorder_sum_all_time: r = 0.89
   • days_until_created_mean_all_time ↔ days_until_firstorder_mean_all_time: r = 0.89
   • days_until_created_mean_all_time ↔ days_until_firstorder_max_all_time: r = 0.89
   • days_until_created_mean_all_time ↔ lastorder_delta_hours_sum_all_time: r = -1.00
   • days_until_created_mean_all_time ↔ lastorder_delta_hours_mean_all_time: r = -1.00
   • days_until_created_mean_all_time ↔ lastorder_delta_hours_max_all_time: r = -1.00
   • days_until_created_mean_all_time ↔ lag0_created_delta_hours_sum: r = 1.00
   • days_until_created_mean_all_time ↔ lag0_created_delta_hours_mean: r = 1.00
   • days_until_created_mean_all_time ↔ lag0_created_delta_hours_max: r = 1.00
   • days_until_created_mean_all_time ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • days_until_created_mean_all_time ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • days_until_created_mean_all_time ↔ created_delta_hours_cohort_zscore: r = 1.00
   • days_until_created_max_all_time ↔ log1p_days_since_created_sum_all_time: r = -0.71
   • days_until_created_max_all_time ↔ log1p_days_since_created_mean_all_time: r = -0.71
   • days_until_created_max_all_time ↔ log1p_days_since_created_max_all_time: r = -0.71
   • days_until_created_max_all_time ↔ days_since_firstorder_sum_all_time: r = -0.89
   • days_until_created_max_all_time ↔ days_since_firstorder_mean_all_time: r = -0.89
   • days_until_created_max_all_time ↔ days_since_firstorder_max_all_time: r = -0.89
   • days_until_created_max_all_time ↔ days_until_firstorder_sum_all_time: r = 0.89
   • days_until_created_max_all_time ↔ days_until_firstorder_mean_all_time: r = 0.89
   • days_until_created_max_all_time ↔ days_until_firstorder_max_all_time: r = 0.89
   • days_until_created_max_all_time ↔ lastorder_delta_hours_sum_all_time: r = -1.00
   • days_until_created_max_all_time ↔ lastorder_delta_hours_mean_all_time: r = -1.00
   • days_until_created_max_all_time ↔ lastorder_delta_hours_max_all_time: r = -1.00
   • days_until_created_max_all_time ↔ lag0_created_delta_hours_sum: r = 1.00
   • days_until_created_max_all_time ↔ lag0_created_delta_hours_mean: r = 1.00
   • days_until_created_max_all_time ↔ lag0_created_delta_hours_max: r = 1.00
   • days_until_created_max_all_time ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • days_until_created_max_all_time ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • days_until_created_max_all_time ↔ created_delta_hours_cohort_zscore: r = 1.00
   • days_until_created_count_all_time ↔ log1p_days_since_created_count_all_time: r = 1.00
   • days_until_created_count_all_time ↔ is_future_created_count_all_time: r = 0.97
   • days_until_created_count_all_time ↔ eopenrate_end: r = 0.94
   • days_until_created_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • days_until_created_count_all_time ↔ avgorder_beginning: r = -0.79
   • days_until_created_count_all_time ↔ avgorder_end: r = -0.79
   • log1p_days_since_created_sum_all_time ↔ log1p_days_since_created_mean_all_time: r = 1.00
   • log1p_days_since_created_sum_all_time ↔ log1p_days_since_created_max_all_time: r = 1.00
   • log1p_days_since_created_sum_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = 0.78
   • log1p_days_since_created_sum_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = 0.78
   • log1p_days_since_created_sum_all_time ↔ log1p_days_since_firstorder_max_all_time: r = 0.78
   • log1p_days_since_created_sum_all_time ↔ lag0_created_delta_hours_sum: r = -0.72
   • log1p_days_since_created_sum_all_time ↔ lag0_created_delta_hours_mean: r = -0.71
   • log1p_days_since_created_sum_all_time ↔ lag0_created_delta_hours_max: r = -0.71
   • log1p_days_since_created_sum_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.72
   • log1p_days_since_created_sum_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.72
   • log1p_days_since_created_sum_all_time ↔ created_delta_hours_cohort_zscore: r = -0.72
   • log1p_days_since_created_mean_all_time ↔ log1p_days_since_created_max_all_time: r = 1.00
   • log1p_days_since_created_mean_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = 0.78
   • log1p_days_since_created_mean_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = 0.78
   • log1p_days_since_created_mean_all_time ↔ log1p_days_since_firstorder_max_all_time: r = 0.78
   • log1p_days_since_created_mean_all_time ↔ lastorder_delta_hours_sum_all_time: r = 0.71
   • log1p_days_since_created_mean_all_time ↔ lastorder_delta_hours_mean_all_time: r = 0.71
   • log1p_days_since_created_mean_all_time ↔ lastorder_delta_hours_max_all_time: r = 0.71
   • log1p_days_since_created_mean_all_time ↔ lag0_created_delta_hours_sum: r = -0.71
   • log1p_days_since_created_mean_all_time ↔ lag0_created_delta_hours_mean: r = -0.71
   • log1p_days_since_created_mean_all_time ↔ lag0_created_delta_hours_max: r = -0.71
   • log1p_days_since_created_mean_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.71
   • log1p_days_since_created_mean_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.71
   • log1p_days_since_created_mean_all_time ↔ created_delta_hours_cohort_zscore: r = -0.71
   • log1p_days_since_created_max_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = 0.78
   • log1p_days_since_created_max_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = 0.78
   • log1p_days_since_created_max_all_time ↔ log1p_days_since_firstorder_max_all_time: r = 0.78
   • log1p_days_since_created_max_all_time ↔ lastorder_delta_hours_sum_all_time: r = 0.71
   • log1p_days_since_created_max_all_time ↔ lastorder_delta_hours_mean_all_time: r = 0.71
   • log1p_days_since_created_max_all_time ↔ lastorder_delta_hours_max_all_time: r = 0.71
   • log1p_days_since_created_max_all_time ↔ lag0_created_delta_hours_sum: r = -0.71
   • log1p_days_since_created_max_all_time ↔ lag0_created_delta_hours_mean: r = -0.71
   • log1p_days_since_created_max_all_time ↔ lag0_created_delta_hours_max: r = -0.71
   • log1p_days_since_created_max_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.71
   • log1p_days_since_created_max_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.71
   • log1p_days_since_created_max_all_time ↔ created_delta_hours_cohort_zscore: r = -0.71
   • log1p_days_since_created_count_all_time ↔ is_future_created_count_all_time: r = 0.97
   • log1p_days_since_created_count_all_time ↔ eopenrate_end: r = 0.94
   • log1p_days_since_created_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • log1p_days_since_created_count_all_time ↔ avgorder_beginning: r = -0.79
   • log1p_days_since_created_count_all_time ↔ avgorder_end: r = -0.79
   • is_missing_created_count_all_time ↔ is_missing_firstorder_count_all_time: r = 1.00
   • is_missing_created_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • is_future_created_count_all_time ↔ eopenrate_end: r = 0.94
   • is_future_created_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • is_future_created_count_all_time ↔ avgorder_beginning: r = -0.79
   • is_future_created_count_all_time ↔ avgorder_end: r = -0.79
   • days_since_firstorder_sum_all_time ↔ days_since_firstorder_mean_all_time: r = 1.00
   • days_since_firstorder_sum_all_time ↔ days_since_firstorder_max_all_time: r = 1.00
   • days_since_firstorder_sum_all_time ↔ days_until_firstorder_sum_all_time: r = -1.00
   • days_since_firstorder_sum_all_time ↔ days_until_firstorder_mean_all_time: r = -1.00
   • days_since_firstorder_sum_all_time ↔ days_until_firstorder_max_all_time: r = -1.00
   • days_since_firstorder_sum_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = 0.75
   • days_since_firstorder_sum_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = 0.75
   • days_since_firstorder_sum_all_time ↔ log1p_days_since_firstorder_max_all_time: r = 0.75
   • days_since_firstorder_sum_all_time ↔ lastorder_delta_hours_sum_all_time: r = 0.84
   • days_since_firstorder_sum_all_time ↔ lastorder_delta_hours_mean_all_time: r = 0.84
   • days_since_firstorder_sum_all_time ↔ lastorder_delta_hours_max_all_time: r = 0.84
   • days_since_firstorder_sum_all_time ↔ lag0_created_delta_hours_sum: r = -0.89
   • days_since_firstorder_sum_all_time ↔ lag0_created_delta_hours_mean: r = -0.89
   • days_since_firstorder_sum_all_time ↔ lag0_created_delta_hours_max: r = -0.89
   • days_since_firstorder_sum_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.89
   • days_since_firstorder_sum_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.89
   • days_since_firstorder_sum_all_time ↔ created_delta_hours_cohort_zscore: r = -0.89
   • days_since_firstorder_mean_all_time ↔ days_since_firstorder_max_all_time: r = 1.00
   • days_since_firstorder_mean_all_time ↔ days_until_firstorder_sum_all_time: r = -1.00
   • days_since_firstorder_mean_all_time ↔ days_until_firstorder_mean_all_time: r = -1.00
   • days_since_firstorder_mean_all_time ↔ days_until_firstorder_max_all_time: r = -1.00
   • days_since_firstorder_mean_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = 0.75
   • days_since_firstorder_mean_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = 0.75
   • days_since_firstorder_mean_all_time ↔ log1p_days_since_firstorder_max_all_time: r = 0.75
   • days_since_firstorder_mean_all_time ↔ lastorder_delta_hours_sum_all_time: r = 0.86
   • days_since_firstorder_mean_all_time ↔ lastorder_delta_hours_mean_all_time: r = 0.86
   • days_since_firstorder_mean_all_time ↔ lastorder_delta_hours_max_all_time: r = 0.86
   • days_since_firstorder_mean_all_time ↔ lag0_created_delta_hours_sum: r = -0.89
   • days_since_firstorder_mean_all_time ↔ lag0_created_delta_hours_mean: r = -0.89
   • days_since_firstorder_mean_all_time ↔ lag0_created_delta_hours_max: r = -0.89
   • days_since_firstorder_mean_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.89
   • days_since_firstorder_mean_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.89
   • days_since_firstorder_mean_all_time ↔ created_delta_hours_cohort_zscore: r = -0.89
   • days_since_firstorder_max_all_time ↔ days_until_firstorder_sum_all_time: r = -1.00
   • days_since_firstorder_max_all_time ↔ days_until_firstorder_mean_all_time: r = -1.00
   • days_since_firstorder_max_all_time ↔ days_until_firstorder_max_all_time: r = -1.00
   • days_since_firstorder_max_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = 0.75
   • days_since_firstorder_max_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = 0.75
   • days_since_firstorder_max_all_time ↔ log1p_days_since_firstorder_max_all_time: r = 0.75
   • days_since_firstorder_max_all_time ↔ lastorder_delta_hours_sum_all_time: r = 0.86
   • days_since_firstorder_max_all_time ↔ lastorder_delta_hours_mean_all_time: r = 0.86
   • days_since_firstorder_max_all_time ↔ lastorder_delta_hours_max_all_time: r = 0.86
   • days_since_firstorder_max_all_time ↔ lag0_created_delta_hours_sum: r = -0.89
   • days_since_firstorder_max_all_time ↔ lag0_created_delta_hours_mean: r = -0.89
   • days_since_firstorder_max_all_time ↔ lag0_created_delta_hours_max: r = -0.89
   • days_since_firstorder_max_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.89
   • days_since_firstorder_max_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.89
   • days_since_firstorder_max_all_time ↔ created_delta_hours_cohort_zscore: r = -0.89
   • days_since_firstorder_count_all_time ↔ days_until_firstorder_count_all_time: r = 1.00
   • days_since_firstorder_count_all_time ↔ log1p_days_since_firstorder_count_all_time: r = 1.00
   • days_since_firstorder_count_all_time ↔ lastorder_delta_hours_count_all_time: r = 0.87
   • days_since_firstorder_count_all_time ↔ lastorder_hour_count_all_time: r = 0.87
   • days_since_firstorder_count_all_time ↔ lastorder_dow_count_all_time: r = 0.87
   • days_since_firstorder_count_all_time ↔ eopenrate_end: r = 0.94
   • days_since_firstorder_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • days_since_firstorder_count_all_time ↔ avgorder_beginning: r = -0.79
   • days_since_firstorder_count_all_time ↔ avgorder_end: r = -0.79
   • days_until_firstorder_sum_all_time ↔ days_until_firstorder_mean_all_time: r = 1.00
   • days_until_firstorder_sum_all_time ↔ days_until_firstorder_max_all_time: r = 1.00
   • days_until_firstorder_sum_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = -0.75
   • days_until_firstorder_sum_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = -0.75
   • days_until_firstorder_sum_all_time ↔ log1p_days_since_firstorder_max_all_time: r = -0.75
   • days_until_firstorder_sum_all_time ↔ lastorder_delta_hours_sum_all_time: r = -0.84
   • days_until_firstorder_sum_all_time ↔ lastorder_delta_hours_mean_all_time: r = -0.84
   • days_until_firstorder_sum_all_time ↔ lastorder_delta_hours_max_all_time: r = -0.84
   • days_until_firstorder_sum_all_time ↔ lag0_created_delta_hours_sum: r = 0.89
   • days_until_firstorder_sum_all_time ↔ lag0_created_delta_hours_mean: r = 0.89
   • days_until_firstorder_sum_all_time ↔ lag0_created_delta_hours_max: r = 0.89
   • days_until_firstorder_sum_all_time ↔ created_delta_hours_vs_cohort_mean: r = 0.89
   • days_until_firstorder_sum_all_time ↔ created_delta_hours_vs_cohort_pct: r = -0.89
   • days_until_firstorder_sum_all_time ↔ created_delta_hours_cohort_zscore: r = 0.89
   • days_until_firstorder_mean_all_time ↔ days_until_firstorder_max_all_time: r = 1.00
   • days_until_firstorder_mean_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = -0.75
   • days_until_firstorder_mean_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = -0.75
   • days_until_firstorder_mean_all_time ↔ log1p_days_since_firstorder_max_all_time: r = -0.75
   • days_until_firstorder_mean_all_time ↔ lastorder_delta_hours_sum_all_time: r = -0.86
   • days_until_firstorder_mean_all_time ↔ lastorder_delta_hours_mean_all_time: r = -0.86
   • days_until_firstorder_mean_all_time ↔ lastorder_delta_hours_max_all_time: r = -0.86
   • days_until_firstorder_mean_all_time ↔ lag0_created_delta_hours_sum: r = 0.89
   • days_until_firstorder_mean_all_time ↔ lag0_created_delta_hours_mean: r = 0.89
   • days_until_firstorder_mean_all_time ↔ lag0_created_delta_hours_max: r = 0.89
   • days_until_firstorder_mean_all_time ↔ created_delta_hours_vs_cohort_mean: r = 0.89
   • days_until_firstorder_mean_all_time ↔ created_delta_hours_vs_cohort_pct: r = -0.89
   • days_until_firstorder_mean_all_time ↔ created_delta_hours_cohort_zscore: r = 0.89
   • days_until_firstorder_max_all_time ↔ log1p_days_since_firstorder_sum_all_time: r = -0.75
   • days_until_firstorder_max_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = -0.75
   • days_until_firstorder_max_all_time ↔ log1p_days_since_firstorder_max_all_time: r = -0.75
   • days_until_firstorder_max_all_time ↔ lastorder_delta_hours_sum_all_time: r = -0.86
   • days_until_firstorder_max_all_time ↔ lastorder_delta_hours_mean_all_time: r = -0.86
   • days_until_firstorder_max_all_time ↔ lastorder_delta_hours_max_all_time: r = -0.86
   • days_until_firstorder_max_all_time ↔ lag0_created_delta_hours_sum: r = 0.89
   • days_until_firstorder_max_all_time ↔ lag0_created_delta_hours_mean: r = 0.89
   • days_until_firstorder_max_all_time ↔ lag0_created_delta_hours_max: r = 0.89
   • days_until_firstorder_max_all_time ↔ created_delta_hours_vs_cohort_mean: r = 0.89
   • days_until_firstorder_max_all_time ↔ created_delta_hours_vs_cohort_pct: r = -0.89
   • days_until_firstorder_max_all_time ↔ created_delta_hours_cohort_zscore: r = 0.89
   • days_until_firstorder_count_all_time ↔ log1p_days_since_firstorder_count_all_time: r = 1.00
   • days_until_firstorder_count_all_time ↔ lastorder_delta_hours_count_all_time: r = 0.87
   • days_until_firstorder_count_all_time ↔ lastorder_hour_count_all_time: r = 0.87
   • days_until_firstorder_count_all_time ↔ lastorder_dow_count_all_time: r = 0.87
   • days_until_firstorder_count_all_time ↔ eopenrate_end: r = 0.94
   • days_until_firstorder_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • days_until_firstorder_count_all_time ↔ avgorder_beginning: r = -0.79
   • days_until_firstorder_count_all_time ↔ avgorder_end: r = -0.79
   • log1p_days_since_firstorder_sum_all_time ↔ log1p_days_since_firstorder_mean_all_time: r = 1.00
   • log1p_days_since_firstorder_sum_all_time ↔ log1p_days_since_firstorder_max_all_time: r = 1.00
   • log1p_days_since_firstorder_mean_all_time ↔ log1p_days_since_firstorder_max_all_time: r = 1.00
   • log1p_days_since_firstorder_count_all_time ↔ lastorder_delta_hours_count_all_time: r = 0.87
   • log1p_days_since_firstorder_count_all_time ↔ lastorder_hour_count_all_time: r = 0.87
   • log1p_days_since_firstorder_count_all_time ↔ lastorder_dow_count_all_time: r = 0.87
   • log1p_days_since_firstorder_count_all_time ↔ eopenrate_end: r = 0.94
   • log1p_days_since_firstorder_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • log1p_days_since_firstorder_count_all_time ↔ avgorder_beginning: r = -0.79
   • log1p_days_since_firstorder_count_all_time ↔ avgorder_end: r = -0.79
   • is_missing_firstorder_count_all_time ↔ lastorder_is_weekend_count_all_time: r = 1.00
   • is_future_firstorder_count_all_time ↔ eopenrate_end: r = 0.94
   • is_future_firstorder_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • is_future_firstorder_count_all_time ↔ avgorder_beginning: r = -0.79
   • is_future_firstorder_count_all_time ↔ avgorder_end: r = -0.79
   • lastorder_delta_hours_sum_all_time ↔ lastorder_delta_hours_mean_all_time: r = 1.00
   • lastorder_delta_hours_sum_all_time ↔ lastorder_delta_hours_max_all_time: r = 1.00
   • lastorder_delta_hours_sum_all_time ↔ lag0_created_delta_hours_sum: r = -0.95
   • lastorder_delta_hours_sum_all_time ↔ lag0_created_delta_hours_mean: r = -1.00
   • lastorder_delta_hours_sum_all_time ↔ lag0_created_delta_hours_max: r = -1.00
   • lastorder_delta_hours_sum_all_time ↔ eopenrate_end: r = 0.89
   • lastorder_delta_hours_sum_all_time ↔ eopenrate_trend_ratio: r = 0.96
   • lastorder_delta_hours_sum_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.95
   • lastorder_delta_hours_sum_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.95
   • lastorder_delta_hours_sum_all_time ↔ created_delta_hours_cohort_zscore: r = -0.95
   • lastorder_delta_hours_mean_all_time ↔ lastorder_delta_hours_max_all_time: r = 1.00
   • lastorder_delta_hours_mean_all_time ↔ lag0_created_delta_hours_sum: r = -0.95
   • lastorder_delta_hours_mean_all_time ↔ lag0_created_delta_hours_mean: r = -1.00
   • lastorder_delta_hours_mean_all_time ↔ lag0_created_delta_hours_max: r = -1.00
   • lastorder_delta_hours_mean_all_time ↔ eopenrate_end: r = 0.89
   • lastorder_delta_hours_mean_all_time ↔ eopenrate_trend_ratio: r = 0.96
   • lastorder_delta_hours_mean_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.95
   • lastorder_delta_hours_mean_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.95
   • lastorder_delta_hours_mean_all_time ↔ created_delta_hours_cohort_zscore: r = -0.95
   • lastorder_delta_hours_max_all_time ↔ lag0_created_delta_hours_sum: r = -0.95
   • lastorder_delta_hours_max_all_time ↔ lag0_created_delta_hours_mean: r = -1.00
   • lastorder_delta_hours_max_all_time ↔ lag0_created_delta_hours_max: r = -1.00
   • lastorder_delta_hours_max_all_time ↔ eopenrate_end: r = 0.89
   • lastorder_delta_hours_max_all_time ↔ eopenrate_trend_ratio: r = 0.96
   • lastorder_delta_hours_max_all_time ↔ created_delta_hours_vs_cohort_mean: r = -0.95
   • lastorder_delta_hours_max_all_time ↔ created_delta_hours_vs_cohort_pct: r = 0.95
   • lastorder_delta_hours_max_all_time ↔ created_delta_hours_cohort_zscore: r = -0.95
   • lastorder_delta_hours_count_all_time ↔ lastorder_hour_count_all_time: r = 1.00
   • lastorder_delta_hours_count_all_time ↔ lastorder_dow_count_all_time: r = 1.00
   • lastorder_delta_hours_count_all_time ↔ eopenrate_end: r = 0.94
   • lastorder_delta_hours_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • lastorder_delta_hours_count_all_time ↔ avgorder_beginning: r = -0.79
   • lastorder_delta_hours_count_all_time ↔ avgorder_end: r = -0.79
   • lastorder_hour_count_all_time ↔ lastorder_dow_count_all_time: r = 1.00
   • lastorder_hour_count_all_time ↔ eopenrate_end: r = 0.94
   • lastorder_hour_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • lastorder_hour_count_all_time ↔ avgorder_beginning: r = -0.79
   • lastorder_hour_count_all_time ↔ avgorder_end: r = -0.79
   • lastorder_dow_sum_all_time ↔ lastorder_dow_mean_all_time: r = 1.00
   • lastorder_dow_sum_all_time ↔ lastorder_dow_max_all_time: r = 1.00
   • lastorder_dow_sum_all_time ↔ eopenrate_end: r = 0.94
   • lastorder_dow_sum_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • lastorder_dow_sum_all_time ↔ avgorder_beginning: r = -0.79
   • lastorder_dow_sum_all_time ↔ avgorder_end: r = -0.79
   • lastorder_dow_mean_all_time ↔ lastorder_dow_max_all_time: r = 1.00
   • lastorder_dow_mean_all_time ↔ eopenrate_end: r = -0.94
   • lastorder_dow_mean_all_time ↔ eopenrate_trend_ratio: r = -0.99
   • lastorder_dow_mean_all_time ↔ avgorder_beginning: r = 0.79
   • lastorder_dow_mean_all_time ↔ avgorder_end: r = 0.79
   • lastorder_dow_count_all_time ↔ eopenrate_end: r = 0.94
   • lastorder_dow_count_all_time ↔ eopenrate_trend_ratio: r = 0.99
   • lastorder_dow_count_all_time ↔ avgorder_beginning: r = -0.79
   • lastorder_dow_count_all_time ↔ avgorder_end: r = -0.79
   • days_since_first_event_x ↔ cohort_year: r = -0.99
   • days_since_first_event_x ↔ esent_trend_ratio: r = -0.71
   • days_since_first_event_x ↔ eopenrate_end: r = -0.74
   • days_since_first_event_x ↔ avgorder_beginning: r = 0.96
   • days_since_first_event_x ↔ avgorder_end: r = 0.96
   • dow_cos ↔ esent_trend_ratio: r = -0.82
   • cohort_year ↔ esent_trend_ratio: r = 0.75
   • cohort_year ↔ avgorder_beginning: r = -0.94
   • cohort_year ↔ avgorder_end: r = -0.94
   • lag0_esent_sum ↔ lag0_esent_mean: r = 1.00
   • lag0_esent_sum ↔ lag0_esent_max: r = 1.00
   • lag0_esent_sum ↔ esent_end: r = 1.00
   • lag0_esent_sum ↔ esent_trend_ratio: r = 0.78
   • lag0_esent_sum ↔ esent_vs_cohort_mean: r = 1.00
   • lag0_esent_sum ↔ esent_vs_cohort_pct: r = 1.00
   • lag0_esent_sum ↔ esent_cohort_zscore: r = 1.00
   • lag0_esent_mean ↔ lag0_esent_max: r = 1.00
   • lag0_esent_mean ↔ esent_end: r = 1.00
   • lag0_esent_mean ↔ esent_trend_ratio: r = 0.78
   • lag0_esent_mean ↔ esent_vs_cohort_mean: r = 1.00
   • lag0_esent_mean ↔ esent_vs_cohort_pct: r = 1.00
   • lag0_esent_mean ↔ esent_cohort_zscore: r = 1.00
   • lag0_esent_max ↔ esent_end: r = 1.00
   • lag0_esent_max ↔ esent_trend_ratio: r = 0.78
   • lag0_esent_max ↔ esent_vs_cohort_mean: r = 1.00
   • lag0_esent_max ↔ esent_vs_cohort_pct: r = 1.00
   • lag0_esent_max ↔ esent_cohort_zscore: r = 1.00
   • lag0_eopenrate_sum ↔ lag0_eopenrate_mean: r = 1.00
   • lag0_eopenrate_sum ↔ lag0_eopenrate_max: r = 1.00
   • lag0_eopenrate_sum ↔ eopenrate_end: r = 1.00
   • lag0_eopenrate_sum ↔ eopenrate_trend_ratio: r = 0.97
   • lag0_eopenrate_sum ↔ avgorder_beginning: r = -0.85
   • lag0_eopenrate_sum ↔ avgorder_end: r = -0.85
   • lag0_eopenrate_sum ↔ eopenrate_vs_cohort_mean: r = 1.00
   • lag0_eopenrate_sum ↔ eopenrate_vs_cohort_pct: r = 1.00
   • lag0_eopenrate_sum ↔ eopenrate_cohort_zscore: r = 1.00
   • lag0_eopenrate_mean ↔ lag0_eopenrate_max: r = 1.00
   • lag0_eopenrate_mean ↔ eopenrate_end: r = 1.00
   • lag0_eopenrate_mean ↔ eopenrate_trend_ratio: r = 0.97
   • lag0_eopenrate_mean ↔ avgorder_beginning: r = -0.85
   • lag0_eopenrate_mean ↔ avgorder_end: r = -0.85
   • lag0_eopenrate_mean ↔ eopenrate_vs_cohort_mean: r = 1.00
   • lag0_eopenrate_mean ↔ eopenrate_vs_cohort_pct: r = 1.00
   • lag0_eopenrate_mean ↔ eopenrate_cohort_zscore: r = 1.00
   • lag0_eopenrate_max ↔ eopenrate_end: r = 1.00
   • lag0_eopenrate_max ↔ eopenrate_trend_ratio: r = 0.97
   • lag0_eopenrate_max ↔ avgorder_beginning: r = -0.85
   • lag0_eopenrate_max ↔ avgorder_end: r = -0.85
   • lag0_eopenrate_max ↔ eopenrate_vs_cohort_mean: r = 1.00
   • lag0_eopenrate_max ↔ eopenrate_vs_cohort_pct: r = 1.00
   • lag0_eopenrate_max ↔ eopenrate_cohort_zscore: r = 1.00
   • lag0_eclickrate_sum ↔ lag0_eclickrate_mean: r = 1.00
   • lag0_eclickrate_sum ↔ lag0_eclickrate_max: r = 1.00
   • lag0_eclickrate_sum ↔ eopenrate_end: r = 0.94
   • lag0_eclickrate_sum ↔ eopenrate_trend_ratio: r = 0.99
   • lag0_eclickrate_sum ↔ avgorder_beginning: r = -0.79
   • lag0_eclickrate_sum ↔ avgorder_end: r = -0.79
   • lag0_eclickrate_sum ↔ eclickrate_vs_cohort_mean: r = 1.00
   • lag0_eclickrate_sum ↔ eclickrate_vs_cohort_pct: r = 1.00
   • lag0_eclickrate_sum ↔ eclickrate_cohort_zscore: r = 1.00
   • lag0_eclickrate_mean ↔ lag0_eclickrate_max: r = 1.00
   • lag0_eclickrate_mean ↔ eopenrate_end: r = 0.94
   • lag0_eclickrate_mean ↔ eopenrate_trend_ratio: r = 0.99
   • lag0_eclickrate_mean ↔ avgorder_beginning: r = -0.79
   • lag0_eclickrate_mean ↔ avgorder_end: r = -0.79
   • lag0_eclickrate_mean ↔ eclickrate_vs_cohort_mean: r = 1.00
   • lag0_eclickrate_mean ↔ eclickrate_vs_cohort_pct: r = 1.00
   • lag0_eclickrate_mean ↔ eclickrate_cohort_zscore: r = 1.00
   • lag0_eclickrate_max ↔ eopenrate_end: r = 0.94
   • lag0_eclickrate_max ↔ eopenrate_trend_ratio: r = 0.99
   • lag0_eclickrate_max ↔ avgorder_beginning: r = -0.79
   • lag0_eclickrate_max ↔ avgorder_end: r = -0.79
   • lag0_eclickrate_max ↔ eclickrate_vs_cohort_mean: r = 1.00
   • lag0_eclickrate_max ↔ eclickrate_vs_cohort_pct: r = 1.00
   • lag0_eclickrate_max ↔ eclickrate_cohort_zscore: r = 1.00
   • lag0_avgorder_sum ↔ lag0_avgorder_mean: r = 1.00
   • lag0_avgorder_sum ↔ lag0_avgorder_max: r = 1.00
   • lag0_avgorder_sum ↔ esent_trend_ratio: r = -0.76
   • lag0_avgorder_sum ↔ eopenrate_end: r = -0.85
   • lag0_avgorder_sum ↔ eopenrate_trend_ratio: r = -0.87
   • lag0_avgorder_sum ↔ avgorder_beginning: r = 1.00
   • lag0_avgorder_sum ↔ avgorder_end: r = 1.00
   • lag0_avgorder_sum ↔ avgorder_vs_cohort_mean: r = 1.00
   • lag0_avgorder_sum ↔ avgorder_vs_cohort_pct: r = 1.00
   • lag0_avgorder_sum ↔ avgorder_cohort_zscore: r = 1.00
   • lag0_avgorder_mean ↔ lag0_avgorder_max: r = 1.00
   • lag0_avgorder_mean ↔ esent_trend_ratio: r = -0.76
   • lag0_avgorder_mean ↔ eopenrate_end: r = -0.85
   • lag0_avgorder_mean ↔ eopenrate_trend_ratio: r = -0.87
   • lag0_avgorder_mean ↔ avgorder_beginning: r = 1.00
   • lag0_avgorder_mean ↔ avgorder_end: r = 1.00
   • lag0_avgorder_mean ↔ avgorder_vs_cohort_mean: r = 1.00
   • lag0_avgorder_mean ↔ avgorder_vs_cohort_pct: r = 1.00
   • lag0_avgorder_mean ↔ avgorder_cohort_zscore: r = 1.00
   • lag0_avgorder_max ↔ esent_trend_ratio: r = -0.76
   • lag0_avgorder_max ↔ eopenrate_end: r = -0.85
   • lag0_avgorder_max ↔ eopenrate_trend_ratio: r = -0.87
   • lag0_avgorder_max ↔ avgorder_beginning: r = 1.00
   • lag0_avgorder_max ↔ avgorder_end: r = 1.00
   • lag0_avgorder_max ↔ avgorder_vs_cohort_mean: r = 1.00
   • lag0_avgorder_max ↔ avgorder_vs_cohort_pct: r = 1.00
   • lag0_avgorder_max ↔ avgorder_cohort_zscore: r = 1.00
   • lag0_ordfreq_sum ↔ lag0_ordfreq_mean: r = 1.00
   • lag0_ordfreq_sum ↔ lag0_ordfreq_max: r = 1.00
   • lag0_ordfreq_sum ↔ esent_beginning: r = 0.71
   • lag0_ordfreq_sum ↔ ordfreq_vs_cohort_mean: r = 1.00
   • lag0_ordfreq_sum ↔ ordfreq_vs_cohort_pct: r = 1.00
   • lag0_ordfreq_sum ↔ ordfreq_cohort_zscore: r = 1.00
   • lag0_ordfreq_mean ↔ lag0_ordfreq_max: r = 1.00
   • lag0_ordfreq_mean ↔ esent_beginning: r = 0.71
   • lag0_ordfreq_mean ↔ ordfreq_vs_cohort_mean: r = 1.00
   • lag0_ordfreq_mean ↔ ordfreq_vs_cohort_pct: r = 1.00
   • lag0_ordfreq_mean ↔ ordfreq_cohort_zscore: r = 1.00
   • lag0_ordfreq_max ↔ esent_beginning: r = 0.71
   • lag0_ordfreq_max ↔ ordfreq_vs_cohort_mean: r = 1.00
   • lag0_ordfreq_max ↔ ordfreq_vs_cohort_pct: r = 1.00
   • lag0_ordfreq_max ↔ ordfreq_cohort_zscore: r = 1.00
   • lag0_created_delta_hours_sum ↔ lag0_created_delta_hours_mean: r = 1.00
   • lag0_created_delta_hours_sum ↔ lag0_created_delta_hours_max: r = 1.00
   • lag0_created_delta_hours_sum ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • lag0_created_delta_hours_sum ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • lag0_created_delta_hours_sum ↔ created_delta_hours_cohort_zscore: r = 1.00
   • lag0_created_delta_hours_mean ↔ lag0_created_delta_hours_max: r = 1.00
   • lag0_created_delta_hours_mean ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • lag0_created_delta_hours_mean ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • lag0_created_delta_hours_mean ↔ created_delta_hours_cohort_zscore: r = 1.00
   • lag0_created_delta_hours_max ↔ created_delta_hours_vs_cohort_mean: r = 1.00
   • lag0_created_delta_hours_max ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • lag0_created_delta_hours_max ↔ created_delta_hours_cohort_zscore: r = 1.00
   • esent_beginning ↔ ordfreq_vs_cohort_mean: r = 0.71
   • esent_beginning ↔ ordfreq_vs_cohort_pct: r = 0.71
   • esent_beginning ↔ ordfreq_cohort_zscore: r = 0.71
   • esent_end ↔ esent_trend_ratio: r = 0.78
   • esent_end ↔ esent_vs_cohort_mean: r = 1.00
   • esent_end ↔ esent_vs_cohort_pct: r = 1.00
   • esent_end ↔ esent_cohort_zscore: r = 1.00
   • esent_trend_ratio ↔ eopenrate_trend_ratio: r = 0.75
   • esent_trend_ratio ↔ avgorder_beginning: r = -0.77
   • esent_trend_ratio ↔ avgorder_end: r = -0.76
   • esent_trend_ratio ↔ esent_vs_cohort_mean: r = 0.78
   • esent_trend_ratio ↔ esent_vs_cohort_pct: r = 0.78
   • esent_trend_ratio ↔ esent_cohort_zscore: r = 0.78
   • esent_trend_ratio ↔ avgorder_vs_cohort_mean: r = -0.76
   • esent_trend_ratio ↔ avgorder_vs_cohort_pct: r = -0.76
   • esent_trend_ratio ↔ avgorder_cohort_zscore: r = -0.76
   • eopenrate_beginning ↔ eclickrate_beginning: r = 0.75
   • eopenrate_end ↔ eopenrate_trend_ratio: r = 0.97
   • eopenrate_end ↔ avgorder_beginning: r = -0.85
   • eopenrate_end ↔ avgorder_end: r = -0.85
   • eopenrate_end ↔ eopenrate_vs_cohort_mean: r = 1.00
   • eopenrate_end ↔ eopenrate_vs_cohort_pct: r = 1.00
   • eopenrate_end ↔ eopenrate_cohort_zscore: r = 1.00
   • eopenrate_end ↔ eclickrate_vs_cohort_mean: r = 0.94
   • eopenrate_end ↔ eclickrate_vs_cohort_pct: r = 0.94
   • eopenrate_end ↔ eclickrate_cohort_zscore: r = 0.94
   • eopenrate_end ↔ avgorder_vs_cohort_mean: r = -0.85
   • eopenrate_end ↔ avgorder_vs_cohort_pct: r = -0.85
   • eopenrate_end ↔ avgorder_cohort_zscore: r = -0.85
   • eopenrate_trend_ratio ↔ avgorder_beginning: r = -0.87
   • eopenrate_trend_ratio ↔ avgorder_end: r = -0.87
   • eopenrate_trend_ratio ↔ eopenrate_vs_cohort_mean: r = 0.97
   • eopenrate_trend_ratio ↔ eopenrate_vs_cohort_pct: r = 0.97
   • eopenrate_trend_ratio ↔ eopenrate_cohort_zscore: r = 0.97
   • eopenrate_trend_ratio ↔ eclickrate_vs_cohort_mean: r = 0.99
   • eopenrate_trend_ratio ↔ eclickrate_vs_cohort_pct: r = 0.99
   • eopenrate_trend_ratio ↔ eclickrate_cohort_zscore: r = 0.99
   • eopenrate_trend_ratio ↔ avgorder_vs_cohort_mean: r = -0.87
   • eopenrate_trend_ratio ↔ avgorder_vs_cohort_pct: r = -0.87
   • eopenrate_trend_ratio ↔ avgorder_cohort_zscore: r = -0.87
   • avgorder_beginning ↔ avgorder_end: r = 1.00
   • avgorder_beginning ↔ eopenrate_vs_cohort_mean: r = -0.85
   • avgorder_beginning ↔ eopenrate_vs_cohort_pct: r = -0.85
   • avgorder_beginning ↔ eopenrate_cohort_zscore: r = -0.85
   • avgorder_beginning ↔ eclickrate_vs_cohort_mean: r = -0.79
   • avgorder_beginning ↔ eclickrate_vs_cohort_pct: r = -0.79
   • avgorder_beginning ↔ eclickrate_cohort_zscore: r = -0.79
   • avgorder_beginning ↔ avgorder_vs_cohort_mean: r = 1.00
   • avgorder_beginning ↔ avgorder_vs_cohort_pct: r = 1.00
   • avgorder_beginning ↔ avgorder_cohort_zscore: r = 1.00
   • avgorder_end ↔ eopenrate_vs_cohort_mean: r = -0.85
   • avgorder_end ↔ eopenrate_vs_cohort_pct: r = -0.85
   • avgorder_end ↔ eopenrate_cohort_zscore: r = -0.85
   • avgorder_end ↔ eclickrate_vs_cohort_mean: r = -0.79
   • avgorder_end ↔ eclickrate_vs_cohort_pct: r = -0.79
   • avgorder_end ↔ eclickrate_cohort_zscore: r = -0.79
   • avgorder_end ↔ avgorder_vs_cohort_mean: r = 1.00
   • avgorder_end ↔ avgorder_vs_cohort_pct: r = 1.00
   • avgorder_end ↔ avgorder_cohort_zscore: r = 1.00
   • esent_vs_cohort_mean ↔ esent_vs_cohort_pct: r = 1.00
   • esent_vs_cohort_mean ↔ esent_cohort_zscore: r = 1.00
   • esent_vs_cohort_pct ↔ esent_cohort_zscore: r = 1.00
   • eopenrate_vs_cohort_mean ↔ eopenrate_vs_cohort_pct: r = 1.00
   • eopenrate_vs_cohort_mean ↔ eopenrate_cohort_zscore: r = 1.00
   • eopenrate_vs_cohort_pct ↔ eopenrate_cohort_zscore: r = 1.00
   • eclickrate_vs_cohort_mean ↔ eclickrate_vs_cohort_pct: r = 1.00
   • eclickrate_vs_cohort_mean ↔ eclickrate_cohort_zscore: r = 1.00
   • eclickrate_vs_cohort_pct ↔ eclickrate_cohort_zscore: r = 1.00
   • avgorder_vs_cohort_mean ↔ avgorder_vs_cohort_pct: r = 1.00
   • avgorder_vs_cohort_mean ↔ avgorder_cohort_zscore: r = 1.00
   • avgorder_vs_cohort_pct ↔ avgorder_cohort_zscore: r = 1.00
   • ordfreq_vs_cohort_mean ↔ ordfreq_vs_cohort_pct: r = 1.00
   • ordfreq_vs_cohort_mean ↔ ordfreq_cohort_zscore: r = 1.00
   • ordfreq_vs_cohort_pct ↔ ordfreq_cohort_zscore: r = 1.00
   • created_delta_hours_vs_cohort_mean ↔ created_delta_hours_vs_cohort_pct: r = -1.00
   • created_delta_hours_vs_cohort_mean ↔ created_delta_hours_cohort_zscore: r = 1.00
   • created_delta_hours_vs_cohort_pct ↔ created_delta_hours_cohort_zscore: r = -1.00

   💡 For each pair, keep the feature with:
      - Stronger business meaning
      - Higher target correlation
      - Fewer missing values

----------------------------------------------------------------------
DETAILED RECOMMENDATIONS:

🔴 Remove multicollinear feature
   event_count_all_time and esent_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and eopenrate_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and eclickrate_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and avgorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and refill_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   event_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and esent_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and esent_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and lag0_esent_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and lag0_esent_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and lag0_esent_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and esent_beginning are highly correlated (r=0.90)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and esent_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and esent_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and esent_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_sum_all_time and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and esent_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and lag0_esent_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and lag0_esent_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and lag0_esent_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and esent_beginning are highly correlated (r=0.90)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and esent_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and esent_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and esent_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_mean_all_time and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_max_all_time and lag0_esent_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_max_all_time and lag0_esent_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_max_all_time and lag0_esent_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_max_all_time and esent_beginning are highly correlated (r=0.98)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_max_all_time and esent_end are highly correlated (r=0.82)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_max_all_time and esent_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_max_all_time and esent_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_max_all_time and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and eopenrate_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and eclickrate_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and avgorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and refill_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and eopenrate_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and eopenrate_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and lag0_eopenrate_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and lag0_eopenrate_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and lag0_eopenrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and eopenrate_beginning are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_sum_all_time and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_mean_all_time and eopenrate_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_mean_all_time and lag0_eopenrate_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_mean_all_time and lag0_eopenrate_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_mean_all_time and lag0_eopenrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_mean_all_time and eopenrate_beginning are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_mean_all_time and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_mean_all_time and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_mean_all_time and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_max_all_time and lag0_eopenrate_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_max_all_time and lag0_eopenrate_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_max_all_time and lag0_eopenrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_max_all_time and eopenrate_beginning are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_max_all_time and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_max_all_time and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_max_all_time and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and eclickrate_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and avgorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and refill_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and eclickrate_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and eclickrate_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and lag0_eclickrate_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and lag0_eclickrate_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and lag0_eclickrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and eopenrate_end are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and eopenrate_trend_ratio are highly correlated (r=0.90)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and avgorder_beginning are highly correlated (r=-0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and avgorder_end are highly correlated (r=-0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_sum_all_time and eclickrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and eclickrate_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and lag0_eclickrate_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and lag0_eclickrate_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and lag0_eclickrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and eopenrate_end are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and eopenrate_trend_ratio are highly correlated (r=0.90)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and avgorder_beginning are highly correlated (r=-0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and avgorder_end are highly correlated (r=-0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_mean_all_time and eclickrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and lag0_eclickrate_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and lag0_eclickrate_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and lag0_eclickrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and eopenrate_end are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and eopenrate_trend_ratio are highly correlated (r=0.90)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and avgorder_beginning are highly correlated (r=-0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and avgorder_end are highly correlated (r=-0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_max_all_time and eclickrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and avgorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and refill_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and avgorder_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and avgorder_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and lag0_avgorder_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and lag0_avgorder_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and lag0_avgorder_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_sum_all_time and esent_trend_ratio are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and eopenrate_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and eopenrate_trend_ratio are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and avgorder_beginning are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and avgorder_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and avgorder_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_sum_all_time and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and avgorder_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and lag0_avgorder_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and lag0_avgorder_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and lag0_avgorder_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_mean_all_time and esent_trend_ratio are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and eopenrate_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and eopenrate_trend_ratio are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and avgorder_beginning are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and avgorder_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and avgorder_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_mean_all_time and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and lag0_avgorder_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and lag0_avgorder_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and lag0_avgorder_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_max_all_time and esent_trend_ratio are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and eopenrate_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and eopenrate_trend_ratio are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and avgorder_beginning are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and avgorder_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and avgorder_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_max_all_time and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_count_all_time and ordfreq_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_count_all_time and refill_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_sum_all_time and ordfreq_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_sum_all_time and ordfreq_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_sum_all_time and lag0_ordfreq_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_sum_all_time and lag0_ordfreq_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_sum_all_time and lag0_ordfreq_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_sum_all_time and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_sum_all_time and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_sum_all_time and ordfreq_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_mean_all_time and ordfreq_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_mean_all_time and lag0_ordfreq_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_mean_all_time and lag0_ordfreq_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_mean_all_time and lag0_ordfreq_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_mean_all_time and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_mean_all_time and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_mean_all_time and ordfreq_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_max_all_time and lag0_ordfreq_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_max_all_time and lag0_ordfreq_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_max_all_time and lag0_ordfreq_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_max_all_time and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_max_all_time and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_max_all_time and ordfreq_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_count_all_time and paperless_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_count_all_time and refill_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_sum_all_time and esent_trend_ratio are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_sum_all_time and avgorder_beginning are highly correlated (r=-0.92)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_sum_all_time and avgorder_end are highly correlated (r=-0.92)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_count_all_time and refill_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   paperless_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   refill_count_all_time and doorstep_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   refill_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   refill_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   refill_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   refill_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   doorstep_count_all_time and firstorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   doorstep_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   doorstep_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   doorstep_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and created_delta_hours_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and created_delta_hours_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_since_created_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_since_created_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_since_created_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_until_created_sum_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_until_created_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_sum_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=-0.72)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_sum_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_sum_all_time and log1p_days_since_created_max_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and created_delta_hours_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_since_created_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_since_created_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_since_created_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_until_created_sum_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_until_created_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_mean_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_mean_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_mean_all_time and log1p_days_since_created_max_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_since_created_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_since_created_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_since_created_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_until_created_sum_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_until_created_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_max_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_max_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_max_all_time and log1p_days_since_created_max_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and created_hour_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and created_dow_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and created_is_weekend_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and days_since_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and days_until_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and is_future_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_delta_hours_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_hour_count_all_time and created_dow_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_hour_count_all_time and created_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_hour_count_all_time and days_since_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_hour_count_all_time and days_until_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_hour_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_hour_count_all_time and is_future_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_hour_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_hour_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_hour_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_hour_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_sum_all_time and created_dow_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_sum_all_time and created_dow_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_dow_sum_all_time and dow_sin are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_dow_sum_all_time and esent_beginning are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_mean_all_time and created_dow_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_dow_mean_all_time and dow_sin are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_dow_max_all_time and dow_sin are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_count_all_time and created_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_count_all_time and days_since_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_count_all_time and days_until_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_count_all_time and is_future_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_dow_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_dow_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_dow_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_is_weekend_count_all_time and days_since_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_is_weekend_count_all_time and days_until_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_is_weekend_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_is_weekend_count_all_time and is_future_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_is_weekend_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_is_weekend_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_is_weekend_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   created_is_weekend_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_delta_hours_sum_all_time and firstorder_delta_hours_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_delta_hours_sum_all_time and firstorder_delta_hours_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_delta_hours_sum_all_time and eopenrate_beginning are highly correlated (r=0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_delta_hours_mean_all_time and firstorder_delta_hours_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_delta_hours_mean_all_time and eopenrate_beginning are highly correlated (r=0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_delta_hours_max_all_time and eopenrate_beginning are highly correlated (r=0.93)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_delta_hours_count_all_time and firstorder_hour_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_delta_hours_count_all_time and firstorder_dow_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_delta_hours_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_delta_hours_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_delta_hours_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_hour_count_all_time and firstorder_dow_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_hour_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_hour_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_hour_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_dow_sum_all_time and firstorder_dow_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_dow_sum_all_time and firstorder_dow_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_sum_all_time and firstorder_is_weekend_mean_all_time are highly correlated (r=0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_sum_all_time and eclickrate_beginning are highly correlated (r=0.81)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_dow_mean_all_time and firstorder_dow_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_mean_all_time and firstorder_is_weekend_mean_all_time are highly correlated (r=0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_mean_all_time and eclickrate_beginning are highly correlated (r=0.81)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_max_all_time and firstorder_is_weekend_mean_all_time are highly correlated (r=0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_max_all_time and eclickrate_beginning are highly correlated (r=0.81)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   firstorder_dow_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_is_weekend_count_all_time and is_missing_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_is_weekend_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   firstorder_is_weekend_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_since_created_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_since_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_until_created_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_until_created_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_until_created_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_sum_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=0.72)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_sum_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_sum_all_time and log1p_days_since_created_max_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_since_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_since_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_since_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_since_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_until_created_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_until_created_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_until_created_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_mean_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_mean_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_mean_all_time and log1p_days_since_created_max_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_since_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_since_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_since_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_until_created_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_until_created_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_until_created_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_max_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_max_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_max_all_time and log1p_days_since_created_max_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_since_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_since_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_since_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and days_until_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_count_all_time and days_until_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_count_all_time and is_future_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_created_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_created_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and days_until_created_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_sum_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=-0.72)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_sum_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_sum_all_time and log1p_days_since_created_max_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and days_until_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_mean_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_mean_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_mean_all_time and log1p_days_since_created_max_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_max_all_time and log1p_days_since_created_sum_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_max_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_max_all_time and log1p_days_since_created_max_all_time are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and days_since_firstorder_sum_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and days_since_firstorder_mean_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and days_since_firstorder_max_all_time are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and days_until_firstorder_sum_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and days_until_firstorder_mean_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and days_until_firstorder_max_all_time are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and lag0_created_delta_hours_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_count_all_time and log1p_days_since_created_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_count_all_time and is_future_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_created_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_created_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and log1p_days_since_created_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and log1p_days_since_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.72)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.72)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.72)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.72)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and log1p_days_since_created_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_created_count_all_time and is_future_created_count_all_time are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_created_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_created_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_created_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   is_missing_created_count_all_time and is_missing_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   is_missing_created_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   is_future_created_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   is_future_created_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   is_future_created_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   is_future_created_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and days_since_firstorder_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and days_since_firstorder_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_sum_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_sum_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_sum_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_sum_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.84)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.84)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.84)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and days_since_firstorder_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_mean_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_mean_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_mean_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and days_until_firstorder_sum_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and days_until_firstorder_mean_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and days_until_firstorder_max_all_time are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_max_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_max_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_max_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and lag0_created_delta_hours_max are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_count_all_time and days_until_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_count_all_time and log1p_days_since_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_firstorder_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_firstorder_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_sum_all_time and days_until_firstorder_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_sum_all_time and days_until_firstorder_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_sum_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_sum_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_sum_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_sum_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.84)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.84)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.84)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and days_until_firstorder_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_mean_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_mean_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_mean_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_max_all_time and log1p_days_since_firstorder_sum_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_max_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_max_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=-0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and lastorder_delta_hours_sum_all_time are highly correlated (r=-0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=-0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=-0.86)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and lag0_created_delta_hours_max are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=-0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_count_all_time and log1p_days_since_firstorder_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_until_firstorder_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_until_firstorder_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_firstorder_sum_all_time and log1p_days_since_firstorder_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_firstorder_sum_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_firstorder_mean_all_time and log1p_days_since_firstorder_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_firstorder_count_all_time and lastorder_delta_hours_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_firstorder_count_all_time and lastorder_hour_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_firstorder_count_all_time and lastorder_dow_count_all_time are highly correlated (r=0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_firstorder_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   log1p_days_since_firstorder_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_firstorder_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   log1p_days_since_firstorder_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   is_missing_firstorder_count_all_time and lastorder_is_weekend_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   is_future_firstorder_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   is_future_firstorder_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   is_future_firstorder_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   is_future_firstorder_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and lastorder_delta_hours_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and eopenrate_end are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and eopenrate_trend_ratio are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_sum_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and lastorder_delta_hours_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and eopenrate_end are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and eopenrate_trend_ratio are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_mean_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_max_all_time and lag0_created_delta_hours_sum are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_max_all_time and lag0_created_delta_hours_mean are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_max_all_time and lag0_created_delta_hours_max are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_max_all_time and eopenrate_end are highly correlated (r=0.89)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_max_all_time and eopenrate_trend_ratio are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_max_all_time and created_delta_hours_vs_cohort_mean are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_max_all_time and created_delta_hours_vs_cohort_pct are highly correlated (r=0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_max_all_time and created_delta_hours_cohort_zscore are highly correlated (r=-0.95)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_count_all_time and lastorder_hour_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_count_all_time and lastorder_dow_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_delta_hours_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_delta_hours_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_delta_hours_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_hour_count_all_time and lastorder_dow_count_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_hour_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_hour_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_hour_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_hour_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_sum_all_time and lastorder_dow_mean_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_sum_all_time and lastorder_dow_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_sum_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_sum_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_dow_sum_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_dow_sum_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_mean_all_time and lastorder_dow_max_all_time are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_mean_all_time and eopenrate_end are highly correlated (r=-0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_mean_all_time and eopenrate_trend_ratio are highly correlated (r=-0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_dow_mean_all_time and avgorder_beginning are highly correlated (r=0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_dow_mean_all_time and avgorder_end are highly correlated (r=0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_count_all_time and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lastorder_dow_count_all_time and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_dow_count_all_time and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lastorder_dow_count_all_time and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_first_event_x and cohort_year are highly correlated (r=-0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_first_event_x and esent_trend_ratio are highly correlated (r=-0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   days_since_first_event_x and eopenrate_end are highly correlated (r=-0.74)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_first_event_x and avgorder_beginning are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   days_since_first_event_x and avgorder_end are highly correlated (r=0.96)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   dow_cos and esent_trend_ratio are highly correlated (r=-0.82)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   cohort_year and esent_trend_ratio are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   cohort_year and avgorder_beginning are highly correlated (r=-0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   cohort_year and avgorder_end are highly correlated (r=-0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_sum and lag0_esent_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_sum and lag0_esent_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_sum and esent_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_esent_sum and esent_trend_ratio are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_sum and esent_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_sum and esent_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_sum and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_mean and lag0_esent_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_mean and esent_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_esent_mean and esent_trend_ratio are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_mean and esent_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_mean and esent_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_mean and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_max and esent_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_esent_max and esent_trend_ratio are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_max and esent_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_max and esent_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_esent_max and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and lag0_eopenrate_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and lag0_eopenrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and eopenrate_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and eopenrate_trend_ratio are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and avgorder_beginning are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and avgorder_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_sum and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_mean and lag0_eopenrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_mean and eopenrate_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_mean and eopenrate_trend_ratio are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_mean and avgorder_beginning are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_mean and avgorder_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_mean and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_mean and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_mean and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_max and eopenrate_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_max and eopenrate_trend_ratio are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_max and avgorder_beginning are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_max and avgorder_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_max and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_max and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eopenrate_max and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_sum and lag0_eclickrate_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_sum and lag0_eclickrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_sum and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_sum and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_eclickrate_sum and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_eclickrate_sum and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_sum and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_sum and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_sum and eclickrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_mean and lag0_eclickrate_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_mean and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_mean and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_eclickrate_mean and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_eclickrate_mean and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_mean and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_mean and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_mean and eclickrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_max and eopenrate_end are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_max and eopenrate_trend_ratio are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_eclickrate_max and avgorder_beginning are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_eclickrate_max and avgorder_end are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_max and eclickrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_max and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_eclickrate_max and eclickrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and lag0_avgorder_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and lag0_avgorder_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_avgorder_sum and esent_trend_ratio are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and eopenrate_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and eopenrate_trend_ratio are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and avgorder_beginning are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and avgorder_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and avgorder_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_sum and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_mean and lag0_avgorder_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_avgorder_mean and esent_trend_ratio are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_mean and eopenrate_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_mean and eopenrate_trend_ratio are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_mean and avgorder_beginning are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_mean and avgorder_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_mean and avgorder_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_mean and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_mean and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_avgorder_max and esent_trend_ratio are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_max and eopenrate_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_max and eopenrate_trend_ratio are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_max and avgorder_beginning are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_max and avgorder_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_max and avgorder_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_max and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_avgorder_max and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_sum and lag0_ordfreq_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_sum and lag0_ordfreq_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_ordfreq_sum and esent_beginning are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_sum and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_sum and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_sum and ordfreq_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_mean and lag0_ordfreq_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_ordfreq_mean and esent_beginning are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_mean and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_mean and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_mean and ordfreq_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   lag0_ordfreq_max and esent_beginning are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_max and ordfreq_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_max and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_ordfreq_max and ordfreq_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_sum and lag0_created_delta_hours_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_sum and lag0_created_delta_hours_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_sum and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_sum and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_sum and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_mean and lag0_created_delta_hours_max are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_mean and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_mean and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_mean and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_max and created_delta_hours_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_max and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   lag0_created_delta_hours_max and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_beginning and ordfreq_vs_cohort_mean are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_beginning and ordfreq_vs_cohort_pct are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_beginning and ordfreq_cohort_zscore are highly correlated (r=0.71)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_end and esent_trend_ratio are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_end and esent_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_end and esent_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_end and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and eopenrate_trend_ratio are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and avgorder_beginning are highly correlated (r=-0.77)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and avgorder_end are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and esent_vs_cohort_mean are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and esent_vs_cohort_pct are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and esent_cohort_zscore are highly correlated (r=0.78)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and avgorder_vs_cohort_mean are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and avgorder_vs_cohort_pct are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   esent_trend_ratio and avgorder_cohort_zscore are highly correlated (r=-0.76)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   eopenrate_beginning and eclickrate_beginning are highly correlated (r=0.75)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and eopenrate_trend_ratio are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and avgorder_beginning are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and avgorder_end are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and eopenrate_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and eclickrate_vs_cohort_mean are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and eclickrate_vs_cohort_pct are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and eclickrate_cohort_zscore are highly correlated (r=0.94)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and avgorder_vs_cohort_mean are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and avgorder_vs_cohort_pct are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_end and avgorder_cohort_zscore are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and avgorder_beginning are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and avgorder_end are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and eopenrate_vs_cohort_mean are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and eopenrate_vs_cohort_pct are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and eopenrate_cohort_zscore are highly correlated (r=0.97)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and eclickrate_vs_cohort_mean are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and eclickrate_vs_cohort_pct are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and eclickrate_cohort_zscore are highly correlated (r=0.99)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and avgorder_vs_cohort_mean are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and avgorder_vs_cohort_pct are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_trend_ratio and avgorder_cohort_zscore are highly correlated (r=-0.87)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_beginning and avgorder_end are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_beginning and eopenrate_vs_cohort_mean are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_beginning and eopenrate_vs_cohort_pct are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_beginning and eopenrate_cohort_zscore are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_beginning and eclickrate_vs_cohort_mean are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_beginning and eclickrate_vs_cohort_pct are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_beginning and eclickrate_cohort_zscore are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_beginning and avgorder_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_beginning and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_beginning and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_end and eopenrate_vs_cohort_mean are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_end and eopenrate_vs_cohort_pct are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_end and eopenrate_cohort_zscore are highly correlated (r=-0.85)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_end and eclickrate_vs_cohort_mean are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_end and eclickrate_vs_cohort_pct are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🟡 Remove multicollinear feature
   avgorder_end and eclickrate_cohort_zscore are highly correlated (r=-0.79)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_end and avgorder_vs_cohort_mean are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_end and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_end and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_vs_cohort_mean and esent_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_vs_cohort_mean and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   esent_vs_cohort_pct and esent_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_vs_cohort_mean and eopenrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_vs_cohort_mean and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eopenrate_vs_cohort_pct and eopenrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_vs_cohort_mean and eclickrate_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_vs_cohort_mean and eclickrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   eclickrate_vs_cohort_pct and eclickrate_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_vs_cohort_mean and avgorder_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_vs_cohort_mean and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   avgorder_vs_cohort_pct and avgorder_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_vs_cohort_mean and ordfreq_vs_cohort_pct are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_vs_cohort_mean and ordfreq_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   ordfreq_vs_cohort_pct and ordfreq_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_vs_cohort_mean and created_delta_hours_vs_cohort_pct are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_vs_cohort_mean and created_delta_hours_cohort_zscore are highly correlated (r=1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Remove multicollinear feature
   created_delta_hours_vs_cohort_pct and created_delta_hours_cohort_zscore are highly correlated (r=-1.00)
   → Consider dropping one of these features. Keep the one with stronger business meaning or higher target correlation.

🔴 Prioritize strong predictors
   Top predictive features: esent_max_all_time, esent_mean_all_time, esent_vs_cohort_pct
   → Ensure these features are included in your model and check for data quality issues.

🟢 Consider removing weak predictors
   Features with low predictive power: event_count_all_time, esent_count_all_time, eopenrate_sum_all_time, eopenrate_mean_all_time, eopenrate_max_all_time
   → These features may add noise. Consider removing or combining with other features.

from sklearn.model_selection import train_test_split

# Stratified split by target
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, stratify=y, random_state=42
)

# Multi-column stratification (for categorical segments)
df['stratify_col'] = df['target'].astype(str) + '_' + df['segment']

# Stratification Recommendations
strat_recs = grouped_recs.get(RecommendationCategory.STRATIFICATION, [])

print("=" * 70)
print("STRATIFICATION (Train/Test Split Strategy)")
print("=" * 70)

# High-risk segments
if analysis_summary.high_risk_segments:
    print("\n🎯 HIGH-RISK SEGMENTS (ensure representation in training data):")
    risk_df = pd.DataFrame(analysis_summary.high_risk_segments)
    risk_df["retention_rate"] = risk_df["retention_rate"].apply(lambda x: f"{x:.1%}")
    risk_df["lift"] = risk_df["lift"].apply(lambda x: f"{x:.2f}x")
    display(risk_df[["feature", "segment", "count", "retention_rate", "lift"]])

    print("\n   💡 These segments have below-average retention.")
    print("   → Ensure they're adequately represented in both train and test sets")
    print("   → Consider oversampling or class weights in modeling")

# Display all stratification recommendations
if strat_recs:
    print("\n" + "-" * 70)
    print("STRATIFICATION RECOMMENDATIONS:")
    for rec in strat_recs:
        priority_icon = "🔴" if rec.priority == "high" else "🟡" if rec.priority == "medium" else "🟢"
        print(f"\n{priority_icon} {rec.title}")
        print(f"   {rec.description}")
        print(f"   → {rec.action}")
else:
    print("\n✅ No special stratification requirements detected.")
    print("   Standard stratified split by target variable is sufficient.")

======================================================================
STRATIFICATION (Train/Test Split Strategy)
======================================================================

🎯 HIGH-RISK SEGMENTS (ensure representation in training data):

   💡 These segments have below-average retention.
   → Ensure they're adequately represented in both train and test sets
   → Consider oversampling or class weights in modeling

----------------------------------------------------------------------
STRATIFICATION RECOMMENDATIONS:

🔴 Stratify by cohort_quarter
   Significant variation in retention rates across cohort_quarter categories (spread: 98.9%)
   → Use stratified sampling by cohort_quarter in train/test split to ensure all segments are represented.

🔴 Monitor high-risk segments
   Segments with below-average retention: 2008.0Q3.0, 2009.0Q1.0, 2009.0Q2.0
   → Target these segments for intervention campaigns and ensure adequate representation in training data.

# Model Selection Recommendations
model_recs = grouped_recs.get(RecommendationCategory.MODEL_SELECTION, [])

print("=" * 70)
print("MODEL SELECTION")
print("=" * 70)

if model_recs:
    for rec in model_recs:
        priority_icon = "🔴" if rec.priority == "high" else "🟡" if rec.priority == "medium" else "🟢"
        print(f"\n{priority_icon} {rec.title}")
        print(f"   {rec.description}")
        print(f"   → {rec.action}")

# Summary recommendations based on data characteristics
print("\n" + "-" * 70)
print("RECOMMENDED MODELING APPROACH:")

has_multicollinearity = len(analysis_summary.multicollinear_pairs) > 0
has_strong_linear = len([p for p in analysis_summary.strong_predictors if abs(p.get("effect_size", 0)) >= 0.5]) > 0
has_categoricals = len(categorical_features) > 0

if has_strong_linear and not has_multicollinearity:
    print("\n✅ RECOMMENDED: Start with Logistic Regression")
    print("   • Strong linear relationships detected")
    print("   • Interpretable coefficients for business insights")
    print("   • Fast training and inference")
    print("   • Then compare with tree-based ensemble for potential improvement")
elif has_multicollinearity:
    print("\n✅ RECOMMENDED: Start with Random Forest or XGBoost")
    print("   • Multicollinearity present - tree models handle it naturally")
    print("   • Can keep all features without VIF analysis")
    print("   • Use feature importance to understand contributions")
else:
    print("\n✅ RECOMMENDED: Compare Linear and Tree-Based Models")
    print("   • No clear linear dominance - test both approaches")
    print("   • Logistic Regression for interpretability baseline")
    print("   • Random Forest/XGBoost for potential accuracy gain")

if has_categoricals:
    print("\n💡 CATEGORICAL HANDLING:")
    print("   • For tree models: Consider CatBoost or LightGBM with native categorical support")
    print("   • For linear models: Use target encoding for high-cardinality features")

======================================================================
MODEL SELECTION
======================================================================

🟡 Consider tree-based models for multicollinearity
   Found 927 highly correlated feature pairs
   → Tree-based models (Random Forest, XGBoost) are robust to multicollinearity. For linear models, remove redundant features first.

🟡 Linear models may perform well
   Strong linear relationships detected (avg effect size: 1.35)
   → Start with Logistic Regression as baseline. Clear feature-target relationships suggest interpretable models may work well.

🟡 Categorical features are predictive
   Strong categorical associations: cohort_quarter
   → Use target encoding for tree-based models or one-hot encoding for linear models. Consider CatBoost for native categorical handling.

----------------------------------------------------------------------
RECOMMENDED MODELING APPROACH:

✅ RECOMMENDED: Start with Random Forest or XGBoost
   • Multicollinearity present - tree models handle it naturally
   • Can keep all features without VIF analysis
   • Use feature importance to understand contributions

💡 CATEGORICAL HANDLING:
   • For tree models: Consider CatBoost or LightGBM with native categorical support
   • For linear models: Use target encoding for high-cardinality features

# Feature Engineering Recommendations
eng_recs = grouped_recs.get(RecommendationCategory.FEATURE_ENGINEERING, [])

print("=" * 70)
print("FEATURE ENGINEERING")
print("=" * 70)

if eng_recs:
    for rec in eng_recs:
        priority_icon = "🔴" if rec.priority == "high" else "🟡" if rec.priority == "medium" else "🟢"
        print(f"\n{priority_icon} {rec.title}")
        print(f"   {rec.description}")
        print(f"   → {rec.action}")
        if rec.affected_features:
            print(f"   → Features: {', '.join(rec.affected_features[:5])}")
else:
    print("\n✅ No specific feature engineering recommendations based on correlation patterns.")
    print("   Consider domain-specific features based on business knowledge.")

# Additional suggestions based on strong predictors
if analysis_summary.strong_predictors:
    print("\n" + "-" * 70)
    print("POTENTIAL INTERACTION FEATURES:")
    strong_features = [p["feature"] for p in analysis_summary.strong_predictors[:5]]
    if len(strong_features) >= 2:
        print("\n   Based on strong predictors, consider interactions between:")
        for i, f1 in enumerate(strong_features[:3]):
            for f2 in strong_features[i+1:4]:
                print(f"   • {f1} × {f2}")
        print("\n   💡 Tree-based models discover interactions automatically.")
        print("   → For linear models, create explicit interaction columns.")

======================================================================
FEATURE ENGINEERING
======================================================================

🟢 Consider ratio features
   Moderately correlated pairs may benefit from ratio features: event_count_all_time/firstorder_delta_hours_count_all_time, event_count_all_time/firstorder_hour_count_all_time, event_count_all_time/firstorder_dow_count_all_time
   → Create ratio features (e.g., feature_a / feature_b) to capture relative relationships.
   → Features: event_count_all_time, event_count_all_time, event_count_all_time, firstorder_delta_hours_count_all_time, firstorder_hour_count_all_time

🟢 Test feature interactions
   Interaction terms may capture non-linear relationships
   → Use PolynomialFeatures(interaction_only=True) or tree-based models which automatically discover interactions.
   → Features: event_count_all_time, esent_sum_all_time, esent_mean_all_time, esent_max_all_time

----------------------------------------------------------------------
POTENTIAL INTERACTION FEATURES:

   Based on strong predictors, consider interactions between:
   • esent_sum_all_time × esent_mean_all_time
   • esent_sum_all_time × esent_max_all_time
   • esent_sum_all_time × lag0_esent_sum
   • esent_mean_all_time × esent_max_all_time
   • esent_mean_all_time × lag0_esent_sum
   • esent_max_all_time × lag0_esent_sum

   💡 Tree-based models discover interactions automatically.
   → For linear models, create explicit interaction columns.

# Create summary table of all recommendations
all_recs_data = []
for rec in analysis_summary.recommendations:
    all_recs_data.append({
        "Category": rec.category.value.replace("_", " ").title(),
        "Priority": rec.priority.upper(),
        "Recommendation": rec.title,
        "Action": rec.action[:80] + "..." if len(rec.action) > 80 else rec.action
    })

if all_recs_data:
    recs_df = pd.DataFrame(all_recs_data)

    # Sort by priority
    priority_order = {"HIGH": 0, "MEDIUM": 1, "LOW": 2}
    recs_df["_sort"] = recs_df["Priority"].map(priority_order)
    recs_df = recs_df.sort_values("_sort").drop("_sort", axis=1)

    print("=" * 80)
    print("ALL RECOMMENDATIONS SUMMARY")
    print("=" * 80)
    print(f"\nTotal: {len(recs_df)} recommendations")
    print(f"  🔴 High priority: {len(recs_df[recs_df['Priority'] == 'HIGH'])}")
    print(f"  🟡 Medium priority: {len(recs_df[recs_df['Priority'] == 'MEDIUM'])}")
    print(f"  🟢 Low priority: {len(recs_df[recs_df['Priority'] == 'LOW'])}")

    display(recs_df)

# Save updated findings and recommendations registry
findings.save(FINDINGS_PATH)
registry.save(RECOMMENDATIONS_PATH)

print(f"\n✅ Findings updated with relationship analysis: {FINDINGS_PATH}")
print(f"✅ Recommendations registry saved: {RECOMMENDATIONS_PATH}")
print(f"   Total recommendations in registry: {len(registry.all_recommendations)}")

if _namespace:
    from customer_retention.analysis.auto_explorer.project_context import ProjectContext

    _namespace.merged_dir.mkdir(parents=True, exist_ok=True)
    _all_findings = _namespace.discover_all_findings(prefer_aggregated=True)
    _mgr = ExplorationManager(_namespace.merged_dir, findings_paths=_all_findings)

    _scaffold = []
    if _namespace.project_context_path.exists():
        _ctx = ProjectContext.load(_namespace.project_context_path)
        _scaffold = _ctx.merge_scaffold

    _multi = _mgr.create_multi_dataset_findings(merge_scaffold=_scaffold)
    _multi.save(str(_namespace.multi_dataset_findings_path))
    _per_dataset_recs = []
    for _ds_name in _namespace.list_datasets():
        _ds_dir = _namespace.dataset_findings_dir(_ds_name)
        if _ds_dir.is_dir():
            for _rp in sorted(_ds_dir.glob("*_recommendations.yaml")):
                _per_dataset_recs.append(RecommendationRegistry.load(str(_rp)))
    if _per_dataset_recs:
        _merged = RecommendationRegistry.merge(_per_dataset_recs)
        _merged.save(str(_namespace.merged_recommendations_path))
    print(f"\n✅ Merged findings/recommendations saved to {_namespace.merged_dir}")

================================================================================
ALL RECOMMENDATIONS SUMMARY
================================================================================

Total: 936 recommendations
  🔴 High priority: 750
  🟡 Medium priority: 183
  🟢 Low priority: 3

✅ Findings updated with relationship analysis: /Users/Vital/python/CustomerRetention/experiments/runs/retail-e7471284/datasets/customer_retention_retail/findings/customer_retention_retail_aggregated_findings.yaml
✅ Recommendations registry saved: /Users/Vital/python/CustomerRetention/experiments/runs/retail-e7471284/datasets/customer_retention_retail/findings/customer_retention_retail_aggregated_recommendations.yaml
   Total recommendations in registry: 1138

✅ Merged findings/recommendations saved to /Users/Vital/python/CustomerRetention/experiments/runs/retail-e7471284/merged

Analysis	What It Tells You	Action
High Correlation (r > 0.7)	Features carry redundant information	Consider removing one
Target Correlation	Feature's predictive power	Prioritize high-correlation features
Class Separation	How different retained vs churned look	Good separation = good predictor
Categorical Rates	Retention varies by category	Use for segmentation and encoding

	Mean (Churned)	Mean (Retained)	Median (Churned)	Median (Retained)
Feature
avgorder_beginning	57.920	45.528	57.920	45.750
avgorder_cohort_zscore	-0.008	0.002	-0.264	-0.266
avgorder_count_all_time	1.000	1.000	1.000	1.000
avgorder_end	57.920	45.680	57.920	45.750
avgorder_max_all_time	61.520	61.929	51.020	50.940
...	...	...	...	...
recency_ratio	0.000	0.000	0.000	0.000
refill_beginning	0.000	0.000	0.000	0.000
refill_count_all_time	1.000	1.000	1.000	1.000
refill_end	0.000	0.000	0.000	0.000
regularity_score	1.000	1.000	1.000	1.000

Effect Size	Interpretation	What It Means for Modeling
\|d\| ≥ 0.8	Large	Strong discriminator - prioritize this feature
\|d\| = 0.5-0.8	Medium	Useful predictor - include in model
\|d\| = 0.2-0.5	Small	Weak but may help in combination with others
\|d\| < 0.2	Negligible	Limited predictive value alone

	cohort_quarter	Count	Retention Rate	Lift	% of Data
0	2018.0Q1.0	273	98.9%	1.24x	0.9%
1	2014.0Q1.0	1081	95.9%	1.21x	3.5%
2	2013.0Q4.0	7221	90.5%	1.14x	23.5%
3	2017.0Q4.0	1868	89.7%	1.13x	6.1%
4	2017.0Q2.0	451	83.8%	1.05x	1.5%
5	2013.0Q1.0	947	82.7%	1.04x	3.1%
6	2013.0Q2.0	1925	81.6%	1.03x	6.3%
7	2017.0Q3.0	1178	79.8%	1.00x	3.8%
8	2013.0Q3.0	4998	78.5%	0.99x	16.2%
9	2008.0Q4.0	25	76.0%	0.96x	0.1%
10	2012.0Q2.0	542	75.5%	0.95x	1.8%
11	2017.0Q1.0	228	75.4%	0.95x	0.7%
12	2016.0Q4.0	144	75.0%	0.94x	0.5%
13	2012.0Q1.0	417	73.6%	0.93x	1.4%
14	2016.0Q3.0	392	72.7%	0.91x	1.3%
15	2012.0Q4.0	596	72.1%	0.91x	1.9%
16	2016.0Q2.0	152	71.7%	0.90x	0.5%
17	2016.0Q1.0	118	71.2%	0.90x	0.4%
18	2015.0Q3.0	227	70.5%	0.89x	0.7%
19	2012.0Q3.0	1480	69.9%	0.88x	4.8%
20	2011.0Q4.0	873	69.6%	0.88x	2.8%
21	2015.0Q1.0	158	69.6%	0.88x	0.5%
22	2011.0Q1.0	608	68.8%	0.87x	2.0%
23	2015.0Q4.0	220	67.7%	0.85x	0.7%
24	2011.0Q3.0	844	67.4%	0.85x	2.7%
25	2015.0Q2.0	290	67.2%	0.85x	0.9%
26	2010.0Q4.0	386	67.1%	0.84x	1.3%
27	2011.0Q2.0	1191	65.9%	0.83x	3.9%
28	2014.0Q2.0	63	65.1%	0.82x	0.2%
29	2008.0Q3.0	34	64.7%	0.81x	0.1%
30	2010.0Q3.0	643	64.5%	0.81x	2.1%
31	2009.0Q2.0	114	60.5%	0.76x	0.4%
32	2014.0Q4.0	94	59.6%	0.75x	0.3%
33	2010.0Q2.0	231	58.4%	0.74x	0.8%
34	2014.0Q3.0	167	55.1%	0.69x	0.5%
35	2010.0Q1.0	244	54.9%	0.69x	0.8%
36	2009.0Q3.0	96	50.0%	0.63x	0.3%
37	2009.0Q1.0	24	50.0%	0.63x	0.1%
38	2009.0Q4.0	225	48.4%	0.61x	0.7%

Pattern	What It Means	Action
Linear trend (diagonal line of points)	Strong correlation	Check if redundant; may cause multicollinearity
Curved pattern	Non-linear relationship	Consider polynomial features or transformations
Clusters/groups	Natural segments in data	May benefit from segment-aware modeling
Fan shape (spreading out)	Heteroscedasticity	May need log transform or robust methods
Random scatter	No relationship	Features are independent

Chapter 5: Relationship Analysis¶

Understanding Feature Relationships¶

5.1 Setup¶

5.1b Leakage Exclusion Gate¶

5.2 Numeric Correlation Matrix¶

5.3 High Correlation Pairs¶

5.4 Feature Distributions by Retention Status¶

Interpreting Effect Sizes (Cohen's d)¶

Box Plot Visualization¶

5.5 Feature-Target Correlations¶

5.6 Categorical Feature Analysis¶

5.7 Scatter Plot Matrix (Sample)¶

Interpreting the Scatter Matrix Above¶

5.8 Datetime Feature Analysis¶

5.9 Actionable Recommendations Summary¶

5.9.1 Feature Selection Recommendations¶

5.9.2 Stratification Recommendations¶

5.9.3 Model Selection Recommendations¶

5.9.4 Feature Engineering Recommendations¶

5.9.5 Recommendations Summary Table¶

Summary: What We Learned¶

Analysis Performed¶

Actionable Recommendations Generated¶

Key Metrics Reference¶

Recommended Actions Checklist¶

Next Steps¶

	Column 1	Column 2	Correlation
0	event_count_all_time	esent_count_all_time	1.000
1	event_count_all_time	eopenrate_count_all_time	1.000
2	event_count_all_time	eclickrate_count_all_time	1.000
3	event_count_all_time	avgorder_count_all_time	1.000
4	event_count_all_time	ordfreq_count_all_time	1.000
...	...	...	...
931	ordfreq_vs_cohort_mean	ordfreq_cohort_zscore	1.000
932	ordfreq_vs_cohort_pct	ordfreq_cohort_zscore	1.000
933	created_delta_hours_vs_cohort_mean	created_delta_hours_vs_cohort_pct	-1.000
934	created_delta_hours_vs_cohort_mean	created_delta_hours_cohort_zscore	1.000
935	created_delta_hours_vs_cohort_pct	created_delta_hours_cohort_zscore	-1.000

Feature Type	Example	Typical Insight
Tenure	Days since signup	Longer tenure often = higher retention
Recency	Days since last order	Recent activity = engaged customer
Cohort	Signup month/year	Economic conditions affect cohorts
Day of Week	Signup day	Weekend vs weekday patterns

Category	Purpose	Impact
Feature Selection	Which features to keep/drop	Reduces noise, improves interpretability
Feature Engineering	New features to create	Captures interactions, improves accuracy
Stratification	Train/test split strategy	Ensures fair evaluation, prevents leakage
Model Selection	Which algorithms to try	Matches model to data characteristics

Finding	Linear Models	Tree-Based Models
Strong predictors	Include - will have high coefficients	Include - will appear early in splits
Weak predictors	Consider dropping	May help in interactions
Multicollinear pairs	Drop one feature	Can keep both (trees handle it)

	feature	correlation	effect_size
0	esent_sum_all_time	+0.718	+2.551
5	lag0_esent_max	+0.718	+2.551
15	esent_vs_cohort_pct	+0.718	+2.551
14	esent_vs_cohort_mean	+0.718	+2.551
1	esent_mean_all_time	+0.718	+2.551
16	esent_cohort_zscore	+0.718	+2.551
4	lag0_esent_mean	+0.718	+2.551
3	lag0_esent_sum	+0.718	+2.551
2	esent_max_all_time	+0.718	+2.551
6	esent_beginning	+0.659	+0.000
7	esent_end	+0.660	+0.000
9	eopenrate_end	+0.398	+0.000
10	eopenrate_trend_ratio	+0.433	+0.000
11	eclickrate_beginning	+0.406	+0.000
12	avgorder_beginning	-0.413	+0.000
13	avgorder_end	-0.403	+0.000
8	esent_trend_ratio	+0.658	+0.000

Data Characteristic	Recommended Models	Reason
Strong linear relationships	Logistic Regression, Linear SVM	Interpretable, fast, less overfit risk
Non-linear patterns	Random Forest, XGBoost, LightGBM	Capture complex interactions
High multicollinearity	Tree-based models	Robust to correlated features
Many categorical features	CatBoost, LightGBM	Native categorical handling
Imbalanced classes	Any with class_weight='balanced'	Adjust for minority class

Pattern Found	Feature to Create	Example
Moderate correlation	Ratio feature	`feature_a / feature_b`
Both features predictive	Interaction term	`feature_a * feature_b`
Curved scatter pattern	Polynomial	`feature_a ** 2`
Related semantics	Difference	`total_orders - returned_orders`

	Category	Priority	Recommendation	Action
0	Feature Selection	HIGH	Remove multicollinear feature	Consider dropping one of these features. Keep ...
591	Feature Selection	HIGH	Remove multicollinear feature	Consider dropping one of these features. Keep ...
592	Feature Selection	HIGH	Remove multicollinear feature	Consider dropping one of these features. Keep ...
593	Feature Selection	HIGH	Remove multicollinear feature	Consider dropping one of these features. Keep ...
594	Feature Selection	HIGH	Remove multicollinear feature	Consider dropping one of these features. Keep ...
...	...	...	...	...
572	Feature Selection	MEDIUM	Remove multicollinear feature	Consider dropping one of these features. Keep ...
604	Feature Selection	MEDIUM	Remove multicollinear feature	Consider dropping one of these features. Keep ...
928	Feature Selection	LOW	Consider removing weak predictors	These features may add noise. Consider removin...
934	Feature Engineering	LOW	Consider ratio features	Create ratio features (e.g., feature_a / featu...
935	Feature Engineering	LOW	Test feature interactions	Use PolynomialFeatures(interaction_only=True) ...

Category	What It Tells You	Impact on Pipeline
Feature Selection	Which features to prioritize/drop	Reduces noise, improves interpretability
Stratification	How to split train/test	Ensures fair evaluation
Model Selection	Which algorithms to try first	Matches model to data
Feature Engineering	Interactions to create	Captures non-linear patterns

Data Type	Effect Measure	Strong Signal
Numeric	Cohen's d	\|d\| ≥ 0.8
Numeric	Correlation	\|r\| ≥ 0.5
Categorical	Cramér's V	V ≥ 0.3
Categorical	Lift	< 0.9x or > 1.1x