FilterWrapper<T>

Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.Hybrid

Filter-Wrapper hybrid feature selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

Think of it as a two-stage interview process. First, resumes are quickly screened (filter) to remove obviously unqualified candidates. Then, the remaining candidates go through detailed interviews (wrapper) to select the best ones. This is faster than interviewing everyone.

How It Works

Filter-Wrapper combines the speed of filter methods with the accuracy of wrapper methods. It first uses a filter to reduce the feature space, then applies a wrapper method on the reduced set for fine-tuning.

Constructors

Constructor Summary
FilterWrapper(int nFeaturesToSelect, int filterRatio, bool useForwardSelection, int[] columnIndices)

Properties

Property Summary
Vector<T> FilterScores { get; }
Vector<T> WrapperScores { get; }
bool SupportsInverseTransform { get; }
bool UseForwardSelection { get; }
int FilterRatio { 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)