Module weights 

Class weighting for imbalanced datasets.

Provides the ClassWeight enum and compute_sample_weights function to generate per-sample weights that compensate for class imbalance.

Example

use scry_learn::weights::{ClassWeight, compute_sample_weights};

let targets = vec![0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0];
let weights = compute_sample_weights(&targets, &ClassWeight::Balanced);

// Minority class (1) gets higher weight to compensate for imbalance.
assert!(weights[9] > weights[0]);

Enums

ClassWeight

Strategy for weighting classes during training.

Functions

compute_sample_weights

Compute per-sample weights from target labels and a class weighting strategy.