MutualInfoRegression<T>

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

Mutual Information for regression-based feature selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

While correlation only finds straight-line relationships, mutual information can detect any pattern where knowing one value tells you about another. This is useful for finding features with curved or complex relationships to your target that linear methods would miss.

How It Works

Mutual Information Regression estimates the mutual information between each continuous feature and a continuous target. Unlike correlation, MI can capture non-linear dependencies between variables.

Constructors

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

Properties

Property Summary
Vector<T> MIScores { get; }
bool SupportsInverseTransform { get; }
int NFeaturesToSelect { get; }
int NNeighbors { 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)