RandomProjectionSelector<T>

Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.Projection

Random Projection-based Feature Selection.

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

Implements: IFixedCountFeatureSelector<T>

For Beginners

Random projections preserve important structure in data when projecting to lower dimensions. By looking at which original features contribute most to these projections, we can identify the most important features without expensive computations.

How It Works

Uses random projections to identify features that contribute most to preserving distances in the projected space (Johnson-Lindenstrauss lemma).

Constructors

Constructor Summary
RandomProjectionSelector(int nFeaturesToSelect, int nProjections, int projectionDim, int[] columnIndices)

Properties

Property Summary
Vector<T> FeatureScores { get; }
bool SupportsInverseTransform { get; }
int NFeaturesToSelect { get; }
int[] SelectedIndices { get; }

Methods

Method Summary
bool[] GetSupportMask()
string[] GetFeatureNamesOut(string[] inputFeatureNames)