Module imbalance 

Imbalanced dataset detection for ML pipelines

Detects class imbalance in classification datasets and provides recommendations for handling strategies.

Example

use alimentar::imbalance::ImbalanceDetector;

let detector = ImbalanceDetector::new("label");
let report = detector.analyze(&dataset)?;

if report.is_imbalanced() {
    println!("Imbalance ratio: {:.2}", report.metrics.imbalance_ratio);
    for rec in &report.recommendations {
        println!("Recommendation: {}", rec);
    }
}

Structs

ClassDistribution

Distribution of classes in a dataset

ImbalanceDetector

Detector for class imbalance in datasets

ImbalanceMetrics

Metrics for measuring class imbalance

ImbalanceReport

Report from imbalance analysis

Enums

ImbalanceRecommendation

Recommendation for handling imbalanced data

ImbalanceSeverity

Severity of class imbalance

ResampleStrategy

Strategy for resampling an imbalanced dataset.

Functions

resample

Resample a classification dataset to address class imbalance.

sqrt_inverse_weights

Compute sqrt-inverse class weights for weighted loss.