MMDBasedSelector<T>

Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.Kernel

Maximum Mean Discrepancy (MMD) based Feature Selection.

Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> โ†’ MMDBasedSelector<T>

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

MMD measures how different two probability distributions are. By computing MMD between different classes, we can find features that make the classes most distinguishable. Features with high MMD are good at separating classes.

How It Works

Uses MMD to measure the distribution difference between classes in kernel space, selecting features that maximize this discrepancy.

Constructors

Constructor Summary
MMDBasedSelector(int nFeaturesToSelect, double gamma, int[] columnIndices)

Properties

Property Summary
Vector<T> MMDScores { get; }
bool SupportsInverseTransform { get; }
double Gamma { 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)