Skip to main content

class_counts

dataset_ml::preprocessing

Function class_counts 

Source 
pub fn class_counts<T: Clone + Ord>(labels: &Array1<T>) -> Vec<(T, usize)>
Expand description

Count how many samples carry each label.

This is a quick way to see how balanced a dataset is before choosing between train_test_split and stratified_split. This function returns counts in sorted class order, matching the numbering label_encode assigns.

§Parameters

  • labels - The per-sample labels to count.

§Returns

  • Vec<(T, usize)> - Each distinct class and its sample count, sorted by class.

§Example

use dataset_ml::preprocessing::class_counts;
use ndarray::array;

let labels = array!["spam", "ham", "ham", "ham"];
assert_eq!(class_counts(&labels), vec![("ham", 3), ("spam", 1)]);