MRMR<T>

Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.Filter.InformationTheory

Minimum Redundancy Maximum Relevance (mRMR) for feature selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

mRMR tries to find features that are both useful and non-redundant. If two features tell you basically the same thing, you only need one of them. This method picks features that give you new, useful information rather than repeating what you already know.

How It Works

mRMR selects features that have high mutual information with the target (relevance) while having low mutual information with already selected features (redundancy).

Constructors

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

Properties

Property Summary
Vector<T> MRMRScores { get; }
bool SupportsInverseTransform { get; }
int NBins { 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)