holisticai.bias.mitigation.DisparateImpactRemoverRS

class holisticai.bias.mitigation.DisparateImpactRemoverRS(*args, **kargs)

Disparate impact remover edits feature values to increase group fairness while preserving rank-ordering within groups.

Parameters

group_colstr

Name of the column in data that contains protected attribute.

score_colstr

Name of the column in data that contains judgment values.

repair_levelfloat

Repair amount 0.0 (min) -> 1.0 (max)

verboseint

If > 0, print progress.

Examples

>>> from holisticai.bias.mitigation import DisparateImpactRemoverRS
>>> mitigator = DisparateImpactRemoverRS(**params)
>>> new_rankings = mitigator.transform(rankings)

References

[1]

Feldman, Michael, et al. “Certifying and removing disparate impact.” proceedings of the 21th ACM SIGKDD international conference on knowledge discovery and data mining. 2015.

fit()

Fit the model

Returns

self

fit_transform(rankings: DataFrame)

Train a Disparate Exposure model to rank the prediction set, then transform the data.

Parameters

rankingsDataFrame

The input data

Returns

DataFrame

Transformed data

transform(rankings: DataFrame)

Train a Disparate Exposure model to rank the prediction set.

Parameters

rankingsDataFrame

The input data

Returns

DataFrame

Transformed data

transform_features(ranking: DataFrame)

Transform data

Description

Transform data to a fair representation

Parameters

rankingDataFrame

Input data

Returns

DataFrame

Transformed data