WinsorizedSelector<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.Robust

Winsorized Feature Selection.

Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> → WinsorizedSelector<T>

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

Winsorizing means replacing the most extreme values in your data with less extreme ones (typically at the 5th and 95th percentile). This way, outliers don’t disappear but are “capped” at reasonable values, giving more robust feature importance measurements.

How It Works

Uses Winsorization to replace extreme values with less extreme values before computing feature importance, providing robust feature selection.

Constructors

Constructor Summary
WinsorizedSelector(int nFeaturesToSelect, double lowerPercentile, double upperPercentile, int[] columnIndices)

Properties

Property Summary
Vector<T> WinsorizedScores { get; }
bool SupportsInverseTransform { get; }
double LowerPercentile { get; }
double UpperPercentile { 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)