Bias Detection¶
Interactive Tool
Quick Reference
- Metrics: Demographic parity, disparity ratio, 80% rule compliance
- Input: CSV with predictions and protected attributes
- Output: Bias report with remediation suggestions
- Related: Bias Testing Guide
Overview¶
The Bias Detection Tool analyzes AI model predictions for fairness across demographic groups:
| Metric | Description |
|---|---|
| Demographic Parity | Equal positive prediction rates |
| Disparity Ratio | Ratio between group rates |
| 80% Rule | EEOC compliance threshold |
| Statistical Parity | Distribution equality |
Features¶
- Upload prediction data
- Multiple protected attribute analysis
- Disparity ratio calculation
- 80% rule compliance check
- Visual bias reporting
- Remediation suggestions
Bias Metrics Explained¶
Demographic Parity¶
Positive outcomes should be equally distributed across groups, regardless of the protected attribute.
80% Rule¶
The selection rate for any protected group should be at least 80% of the rate for the highest group (EEOC guideline).
Disparity Ratio¶
The ratio of positive prediction rates between groups. Values close to 1.0 indicate parity.
When to Use¶
- Model validation before deployment
- Ongoing monitoring
- Compliance reporting
- Audit preparation