Module feature_selection 

Feature selection transformers.

Remove low-information features before training to reduce overfitting and speed up downstream models.

Examples

use scry_learn::prelude::*;

let mut vt = VarianceThreshold::new().threshold(0.1);
vt.fit(&data)?;
vt.transform(&mut data)?; // drops constant / near-constant columns

Structs

SelectKBest

Select the top-k highest-scoring features.

VarianceThreshold

Remove features whose variance falls below a threshold.

Enums

ScoreFn

Scoring function for feature selection.

Functions

f_classif

Compute ANOVA F-value for each feature vs. the target.