LaggedCorrelationSelector<T>

Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.TimeSeries

Lagged correlation selector for time series feature selection.

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

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

This finds features whose current values predict future values of the target. A feature might correlate with the target 3 time steps later, making it useful for forecasting.

How It Works

Computes cross-correlation between features and target at various lags to find features that predict future target values.

Constructors

Constructor Summary
LaggedCorrelationSelector(int nFeaturesToSelect, int minLag, int maxLag, int[] columnIndices)

Properties

Property Summary
Vector<T> CorrelationScores { get; }
bool SupportsInverseTransform { get; }
int NFeaturesToSelect { get; }
int[] BestLags { 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)