Audits a trained classifier or scoring model for disparate treatment across protected attributes (race, sex, age, disability, religion, national origin, sexual orientation, or any user-supplied subgroup) using equal-opportunity, demographic-parity, calibration-within-group, and intersectional cuts.
