AdjustedMutualInformation<T>

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

Adjusted Mutual Information correcting for chance agreement.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

Just by chance, two random groupings will share some information. AMI subtracts this expected chance agreement to give a cleaner measure of true dependency.

How It Works

Adjusted Mutual Information (AMI) corrects for the fact that random labelings have non-zero expected MI. It accounts for the number of clusters to give a more accurate measure of agreement.

Constructors

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

Properties

Property Summary
Vector<T> AMIScores { 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)