SURF<T>

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

Spatially Uniform ReliefF (SURF) for feature selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

Unlike ReliefF which looks at exactly k neighbors, SURF looks at all neighbors within a certain distance. This is more natural because it considers the actual distribution of your data - areas with more data points contribute more comparisons.

How It Works

SURF is a variant of ReliefF that uses distance thresholds instead of k nearest neighbors. It considers all instances within a threshold distance, making it more robust and parameter-free for the neighbor count.

Constructors

Constructor Summary
SURF(int nFeaturesToSelect, double distanceThreshold, int[] columnIndices)

Properties

Property Summary
Vector<T> FeatureWeights { get; }
bool SupportsInverseTransform { get; }
double DistanceThreshold { 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)