holisticai.bias.metrics.cohen_d

holisticai.bias.metrics.cohen_d(group_a, group_b, y_pred)

Cohen D

This function computes the Cohen D statistic (normalised statistical parity) between group_a and group_b.

A value of 0 is desired. Negative values are unfair towards group_a. Positive values are unfair towards group_b. Reference values: 0.2 is considered a small effect size, 0.5 is considered medium, 0.8 is considered large.

group_aarray-like

Group membership vector (binary)

group_barray-like

Group membership vector (binary)

y_predarray-like

Predictions vector (binary)

float

Cohen D :

:math:`

rac{sr_a-sr_b}{ exttt{std_pool}}`

>>> import numpy as np
>>> from holisticai.bias.metrics import cohen_d
>>> group_a = np.array([1, 1, 1, 1, 0, 0, 0, 0, 0, 0])
>>> group_b = np.array([0, 0, 0, 0, 1, 1, 1, 1, 1, 1])
>>> y_pred = np.array([1, 1, 0, 0, 1, 1, 0, 1, 1, 1])
>>> cohen_d(group_a, group_b, y_pred)
-0.7844645405527363