ConcreteAutoencoderSelector<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.Neural

Concrete Autoencoder Feature Selection.

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

Implements: IFixedCountFeatureSelector<T>

For Beginners

The Concrete Autoencoder learns which features to select as part of its training. It uses a special trick (Gumbel-Softmax) that lets it learn a “soft” selection that gradually becomes “hard” (0 or 1) as training progresses.

How It Works

Uses the Concrete Autoencoder which learns a discrete feature selection mask using the Concrete/Gumbel-Softmax distribution for differentiable feature selection.

Constructors

Constructor Summary
ConcreteAutoencoderSelector(int nFeaturesToSelect, int hiddenSize, int nEpochs, double initialTemperature, double minTemperature, int[] columnIndices)

Properties

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

Methods

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