DeepFeatureSelector<T>

Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.Neural

Deep Feature Selection using a multi-layer neural network.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

This method trains a multi-layer neural network but penalizes the input layer weights. Features with larger input weights after training are considered more important. The network can capture complex non-linear patterns that simpler methods might miss.

How It Works

Uses a deep neural network with L1 regularization on the first layer to perform feature selection while learning non-linear relationships.

Constructors

Constructor Summary
DeepFeatureSelector(int nFeaturesToSelect, int[] hiddenLayers, int nEpochs, double l1Penalty, int[] columnIndices)

Properties

Property Summary
Vector<T> FeatureImportances { get; }
bool SupportsInverseTransform { 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)