BimodalitySelector<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.Moment

Bimodality Coefficient based Feature Selection.

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

Implements: IFixedCountFeatureSelector<T>

For Beginners

Bimodality measures if data has two separate groups. A high bimodality coefficient (>0.555 for uniform threshold) suggests the feature naturally separates data into two clusters, which can be useful for classification. The formula is: BC = (skewness² + 1) / (kurtosis + 3 × (n-1)²/((n-2)(n-3))), with the bias-adjusted/n/// sample skewness and excess kurtosis (as in SAS and Pfister et al. 2013).

How It Works

Selects features based on their bimodality coefficient, which indicates whether a distribution has two distinct modes (peaks).

Constructors

Constructor Summary
BimodalitySelector(int nFeaturesToSelect, bool preferHighBimodality, int[] columnIndices)

Properties

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

Methods

Method Summary
Matrix<T> FitTransform(Matrix<T> data)
bool[] GetSupportMask()
string[] GetFeatureNamesOut(string[] inputFeatureNames)