Model: {{ model_name }} ({{ model_type }}) Assessment: {{ assessment }} Protected Attributes: {{ total_attributes }} Warnings: {{ total_warnings }}

Fairness Score {{ "%.4f"|format(overall_fairness_score|default(0)) }}

Protected Attributes {{ total_attributes }}

Warnings {{ total_warnings }}

Critical Issues {{ total_critical }}

Assessment {{ assessment }}

📊 Fairness Assessment Overview

Overall fairness assessment across all protected attributes.

📈 Fairness Metrics Comparison

Comparison of all fairness metrics across protected attributes.

🎯 Fairness Radar

Multi-dimensional fairness profile for each protected attribute.

🔢 Confusion Matrices by Group

🔬 Pre-Training Fairness Analysis

Analysis of fairness in the training data BEFORE model training. These metrics are exclusive to DeepBridge and help identify bias in the data itself.

📊 All Pre-Training Metrics

BCL (Class Balance), BCO (Concept Balance), KL Divergence, JS Divergence

💡 What are Pre-Training Metrics?

BCL (Class Balance): Measures if groups have similar sample sizes
BCO (Concept Balance): Measures if groups have similar positive class rates
KL Divergence: Asymmetric measure of distribution difference
JS Divergence: Symmetric, bounded measure of distribution difference (0-1)

👥 Group Size Distribution

Sample balance across demographic groups - identifies underrepresented groups.

⚖️ Concept Balance

Positive class rate comparison - detects outcome imbalance in training data.

⚖️ Post-Training Detailed Analysis

Advanced fairness metrics after model training, including EEOC compliance monitoring.

🎯 Disparate Impact - EEOC 80% Rule

CRITICAL LEGAL METRIC: Shows compliance with EEOC 80% rule (4/5ths rule).

⚖️ EEOC 80% Rule:

The selection rate for any protected group should be at least 80% of the rate for the highest group. Ratios below 0.8 may indicate adverse impact and potential legal issues.

🟢 ≥0.8: COMPLIANT - Passes EEOC test
🟡 0.7-0.8: WARNING - Borderline compliance
🔴 <0.7: CRITICAL - High legal risk

📊 Statistical Parity - Disparity Analysis

Shows how far each attribute deviates from perfect fairness (0.0 = perfect).

🚦 Compliance Status Matrix

Executive dashboard showing compliance status across all main metrics.

Legend: ✓ Pass ⚠ Warning ✗ Critical

📐 Complementary Fairness Metrics

Additional fairness metrics including exclusive DeepBridge metrics: Treatment Equality and Entropy Index.

🎯 Precision & Accuracy by Group

Performance metrics comparison across demographic groups.

⚖️ Treatment Equality Analysis

EXCLUSIVE DeepBridge metric: Shows if errors (FN vs FP) are balanced across groups.

💡 What is Treatment Equality?

Treatment Equality measures the ratio of False Negatives to False Positives. Groups should have similar error ratios. Points on the diagonal line indicate perfect balance.

🎯 Complementary Metrics Radar

Multi-dimensional view of 6 complementary fairness metrics.

Metrics Included:

Conditional Acceptance: PPV (Positive Predictive Value) parity
Conditional Rejection: NPV (Negative Predictive Value) parity
Precision Difference: Precision gap between groups
Accuracy Difference: Overall accuracy gap
Treatment Equality: FN/FP ratio balance (exclusive)
Entropy Index: Individual fairness via generalized entropy (exclusive)

📊 Data Distribution Analysis

Visualization of data distributions for protected attributes and target variable.

👥 Protected Attributes Distribution

Sample distribution across demographic groups - identifies representation issues.

⚠️ Minimum Representation Threshold:

Groups with less than 2% representation are typically excluded from fairness analysis due to statistical instability (EEOC "Flip-Flop Rule").

🎯 Target Variable Distribution

Distribution of outcomes (classes) in the dataset.

{% if has_threshold_analysis %}

Threshold Analysis

Impact of decision thresholds on fairness metrics.

{% endif %} {% if warnings|length > 0 or critical_issues|length > 0 %}

{% if critical_issues|length > 0 %}

Critical Issues

Issue	Severity	Description

{% endif %} {% if warnings|length > 0 %}

Warnings

Warning	Severity	Description

{% endif %}

{{ report_title }}