holisticai.bias.plots.abroca_plot

holisticai.bias.plots.abroca_plot(group_a, group_b, y_pred, y_true, ax=None, size=None, title=None)

Abroca plot

Description

This function plots the roc curve for both groups revealing the area between them (abroca).

Parameters

group_aarray-like

Group membership vector (binary)

group_barray-like

Group membership vector (binary)

y_predarray-like

Probability estimates (regression)

y_truearray-like

Target vector (binary)

ax (optional)matplotlib axes

Pre-existing axes for the plot

size (optional)(int, int)

Size of the figure

title (optional)str

Title of the figure

Returns

matplotlib ax

Example

>>> from sklearn.linear_model import LogisticRegression
>>> from holisticai.datasets import load_dataset
>>> from holisticai.bias.plots import abroca_plot
>>> X, y = make_classification(n_samples=1000, n_features=20, n_classes=2)
>>> group_a = np.random.randint(0, 2, 1000)
>>> group_b = np.random.randint(0, 2, 1000)
>>> y_pred = LogisticRegression().fit(X, y).predict_proba(X)[:, 1]
>>> abroca_plot(group_a, group_b, y_pred, y)