holisticai.bias.metrics.min_cluster_ratio

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

Minimum Cluster Ratio

Given a clustering and protected attributes. The min cluster ratio is the minimum over all clusters of the ratio of number of group_a members to the number of group_b members.

Interpretation

A value of 1 is desired. That is when all clusters are perfectly balanced. Low values imply the existence of clusters where group_a has fewer members than group_b.

Parameters

group_aarray-like

Group membership vector (binary)

group_barray-like

Group membership vector (binary)

y_predarray-like

Cluster predictions (categorical)

Returns

float

Minimum Cluster Ratio

Examples

>>> import numpy as np
>>> from holisticai.bias.metrics import min_cluster_ratio
>>> 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_cluster = np.array([0, 1, 1, 2, 0, 0, 0, 0, 1, 2])
>>> min_cluster_ratio(group_a, group_b, y_pred_cluster)
0.2499999999375