BorutaFS<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.Ensemble

Boruta algorithm for feature selection.

Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> → BorutaFS<T>

Implements: ISupervisedFeatureSelector<T>

For Beginners

Boruta creates fake versions of your features by randomly shuffling their values. It then asks: “Is this real feature more useful than the best fake feature?” Features that consistently outperform their fake counterparts are kept. This is more rigorous than just picking the top N features.

How It Works

Boruta creates “shadow” features by shuffling each original feature, then trains a model to compare original features against their shadows. Features that consistently beat the best shadow feature are confirmed as important; those that don’t are rejected.

Constructors

Constructor Summary
BorutaFS(int nIterations, double percentile, Func<Matrix<T>, Vector<T>, Vector<T>> getImportances, int? randomState, int[] columnIndices)

Properties

Property Summary
Vector<T> FeatureImportances { get; }
bool SupportsInverseTransform { get; }
double Percentile { get; }
int NIterations { get; }
int[] SelectedIndices { get; }

Methods

Method Summary
Matrix<T> FitTransform(Matrix<T> data, Vector<T> target)
bool[] GetSupportMask()
string[] GetFeatureNamesOut(string[] inputFeatureNames)
void Fit(Matrix<T> data, Vector<T> target)