IterativeStratificationSplitter<T>

Models & Types · AiDotNet.Preprocessing.DataPreparation.Splitting.Stratified

Iterative stratification splitter for multi-label classification problems.

Inheritance: DataSplitterBase<T> → IterativeStratificationSplitter<T>

For Beginners

Standard stratification works well for single-label classification where each sample belongs to exactly one class. But what about multi-label problems where each sample can have multiple labels (like a movie being both “Action” AND “Comedy”)?

How It Works

How It Works:

  1. Calculate label frequencies across all samples
  2. Start with rarest label combinations
  3. Iteratively assign samples to folds, prioritizing balance for rare labels
  4. Continue until all samples are assigned

Example: Document classification where each document can have multiple topics:

  • Doc1: [Politics, Economy]
  • Doc2: [Sports]
  • Doc3: [Politics, Sports]
  • Doc4: [Economy]

Constructors

Constructor Summary
IterativeStratificationSplitter(double testSize, Matrix<T> labelMatrix, bool shuffle, int randomSeed) Creates a new Iterative Stratification splitter for multi-label data.

Properties

Property Summary
bool RequiresLabels { get; }
string Description { get; }

Methods

Method Summary
DataSplitResult<T> Split(Matrix<T> X, Vector<T> y)
IterativeStratificationSplitter<T> WithLabelMatrix(Matrix<T> labelMatrix) Sets the multi-label matrix.