Struct StratifiedCrossValidation 

pub struct StratifiedCrossValidation {
    pub number_of_folds: i64,
    pub random_seed: i64,
    pub scheme_name: String,
    pub description: String,
}

Cross-validation preserving class proportions across folds.

Fields

number_of_folds: i64

Number of folds.

random_seed: i64

Random seed.

scheme_name: String

Unique name identifying a resampling scheme.

description: String

Free-text description.

Implementations

impl StratifiedCrossValidation

pub fn folds(&self, labels: &[usize]) -> Result<Vec<(Vec<usize>, Vec<usize>)>>

Splits labelled observations into stratified folds using this scheme’s configured number_of_folds and random_seed.

Seeds a fresh SplitMix64 from random_seed (reinterpreting the signed seed’s bits as unsigned) and delegates to stratified_kfold_indices, so two calls on structs with equal fields yield identical splits.

Arguments

• labels — the class id of each observation; see stratified_kfold_indices for the partitioning contract.

Returns

A vector of (train_indices, test_indices) pairs, one per fold.

Errors

• Error::InvalidInput if number_of_folds is negative or < 2.
• Error::InsufficientData if labels is empty or number_of_folds exceeds the smallest class count.

Examples

use stats_claw::resampling::StratifiedCrossValidation;

let cv = StratifiedCrossValidation {
    number_of_folds: 2,
    random_seed: 42,
    ..Default::default()
};
let labels = [0usize, 1, 0, 1];
let folds = cv.folds(&labels)?;
assert_eq!(folds.len(), 2);

Trait Implementations

impl Clone for StratifiedCrossValidation

fn clone(&self) -> StratifiedCrossValidation

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for StratifiedCrossValidation

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for StratifiedCrossValidation

fn default() -> StratifiedCrossValidation

Returns the “default value” for a type.