Skip to main content

stratified_sample

Function stratified_sample 

Source
pub fn stratified_sample(
    cases: &[(String, Vec<GoldEntity>)],
    target_size: usize,
    seed: u64,
) -> Vec<(String, Vec<GoldEntity>)>
Expand description

Stratified sampling maintaining entity type proportions.

§Arguments

  • cases - Input test cases (text, gold entities)
  • target_size - Maximum number of cases to return
  • seed - Random seed for reproducibility

§Returns

A subset of cases with proportional entity type distribution.

§Note on True Stratification

This is a simplified version using hash-based pseudo-random ordering. For true stratified sampling by entity type, use stratified_sample_ner which groups by type first.