MultiLabelReliefF<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.MultiLabel

Multi-Label ReliefF for multi-label feature selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

In multi-label problems (like tagging photos with multiple keywords), a single image might be “sunset”, “beach”, and “romantic” all at once. This algorithm finds features that help distinguish between different label combinations, not just individual labels.

How It Works

Multi-Label ReliefF extends the ReliefF algorithm to handle multi-label classification problems where each instance can belong to multiple classes simultaneously. It considers label correlations when computing feature weights.

Constructors

Constructor Summary
MultiLabelReliefF(int nFeaturesToSelect, int nNeighbors, int nIterations, int? randomState, int[] columnIndices)

Properties

Property Summary
Vector<T> FeatureWeights { get; }
bool SupportsInverseTransform { get; }
int NFeaturesToSelect { get; }
int NIterations { get; }
int NNeighbors { get; }
int[] SelectedIndices { get; }

Methods

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