SparseAutoencoderSelector<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.Neural

Sparse Autoencoder-based Feature Selection.

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

Implements: IFixedCountFeatureSelector<T>

For Beginners

A sparse autoencoder forces most of its internal neurons to be “off” most of the time. This makes it focus on the most important input features. Features that activate the sparse neurons are the ones we select.

How It Works

Uses a sparse autoencoder with sparsity constraints on the hidden layer to learn which input features are essential for reconstruction.

Constructors

Constructor Summary
SparseAutoencoderSelector(int nFeaturesToSelect, int hiddenSize, int nEpochs, double sparsityTarget, double sparsityWeight, int[] columnIndices)

Properties

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

Methods

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