error_parity.pareto_curve

Utils for computing the fairness-accuracy Pareto frontier of a classifier.

Functions

compute_inner_and_outer_adjustment_ci(...[, ...])	Computes the interior/inner and exterior/outer adjustment curves, corresponding to the confidence intervals (by default 95% c.i.).
compute_postprocessing_curve(model, ...[, ...])	Computes the fairness and performance of the given classifier after adjusting (postprocessing) for varying levels of fairness tolerance.
fit_and_evaluate_postprocessing(...[, seed, ...])	Fit and evaluate a postprocessing intervention on the given predictor.
get_envelope_of_postprocessing_frontier(...)	Computes points in envelope of the given postprocessing frontier results.