NEWS

---

clinicalfair 0.1.1

• Initial release.
• fairness_data(): Bundle predictions, labels, protected attributes.
• fairness_metrics(): Group-wise fairness metrics (selection rate, TPR, FPR, PPV, accuracy, AUC, Brier) with optional bootstrap CIs.
• fairness_report(): Audit report with four-fifths rule screening.
• threshold_optimize(): Group-specific threshold mitigation (equalized odds / demographic parity) with configurable grid resolution.
• intersectional_fairness(): Cross-tabulated multi-attribute analysis with small-group filtering.
• autoplot(), plot_roc(), plot_calibration(): Visualization.
• Built-in simulated COMPAS dataset for demonstrations.