Module feature_scoring 

Feature scoring functions for feature selection.

This module provides standalone univariate scoring functions that compute per-feature statistics and p-values:

• f_classif — ANOVA F-statistic for classification.
• f_regression — univariate F-statistic via Pearson correlation.
• chi2 — chi-squared statistic for non-negative features.

These functions return (F-statistics, p-values) tuples and can be used directly or passed to SelectKBest / SelectPercentile via the ScoreFunc enum.

Functions

chi2

Compute chi-squared statistics between each non-negative feature and the class labels.

compute_scores_classif

Add FRegression and Chi2 variants to ScoreFunc.

compute_scores_regression

Compute regression scores, returning F-statistics only.

f_classif

Compute the ANOVA F-statistic and approximate p-values for each feature.

f_regression

Compute univariate F-statistics via Pearson correlation for regression.