GenericUnivariateSelect<T>
Models & Types · AiDotNet.Preprocessing.FeatureSelection
Generic univariate feature selector with configurable selection mode.
Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> → GenericUnivariateSelect<T>
Implements: ISupervisedFeatureSelector<T>
For Beginners
This is like a Swiss Army knife for feature selection. Instead of using different classes for different strategies, you pick a mode:
- KBest: “Give me the top 10 features”
- Percentile: “Give me the top 10% of features”
- FPR/FDR/FWE: “Give me features with statistical significance below 0.05”
How It Works
GenericUnivariateSelect is a flexible feature selector that can operate in different modes: k-best, percentile, FPR (false positive rate), FDR (false discovery rate), or FWE (family-wise error rate). This provides a unified interface for various univariate selection strategies.
Constructors
| Constructor | Summary |
|---|---|
GenericUnivariateSelect(SelectionMode mode, double param, Func<Matrix<T>, Vector<T>, ValueTuple<Vector<T>, Vector<T>>> scoreFunc, UnivariateScoreFunction defaultScoreFunc, int[] columnIndices) |
Properties
| Property | Summary |
|---|---|
SelectionMode Mode { get; } |
|
Vector<T> PValues { get; } |
|
Vector<T> Scores { get; } |
|
bool SupportsInverseTransform { get; } |
|
double Param { 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) |