SymmetricUncertainty<T>

Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.Filter

Symmetric Uncertainty Feature Selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

Think of this as measuring how well knowing one value helps predict another. Unlike regular correlation, it works for any type of relationship (not just linear) and always gives a value between 0 and 1, making it easy to compare different features.

How It Works

Symmetric Uncertainty is a normalized version of mutual information that measures the correlation between features and targets. It ranges from 0 (no correlation) to 1 (perfect correlation) and is symmetric.

Constructors

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

Properties

Property Summary
Vector<T> UncertaintyScores { 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)