DISR<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.Filter.Multivariate

Double Input Symmetrical Relevance (DISR) for feature selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

DISR balances relevance and redundancy by looking at how much two features together tell you about the target, normalized by their total information content. It’s like asking: “Of all the information in this feature pair, what fraction is actually useful for prediction?”

How It Works

DISR uses a symmetrical relevance measure that considers both feature-target and feature-feature relationships. It maximizes a normalized joint relevance score that accounts for redundancy in a principled way.

Constructors

Constructor Summary
DISR(int nFeaturesToSelect, int nBins, int[] columnIndices)

Properties

Property Summary
Vector<T> DisrScores { get; }
Vector<T> RelevanceScores { get; }
bool SupportsInverseTransform { get; }
int NFeaturesToSelect { 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)