ScoreCAMSelector<T>

Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.Neural

Score-CAM based Feature Selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

Unlike GradCAM which uses gradients, Score-CAM measures feature importance by masking each feature and observing how the prediction score changes. Features that cause larger score drops when masked are considered more important. This is more stable than gradient methods.

How It Works

Selects features using Score-CAM, a gradient-free class activation mapping method that uses forward passing scores instead of gradients.

Constructors

Constructor Summary
ScoreCAMSelector(int nFeaturesToSelect, int[] columnIndices)

Properties

Property Summary
Vector<T> ScoreWeights { 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)