AutocorrelationFS<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.TimeSeries

Autocorrelation-based feature selection for time series data.

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

Implements: IFixedCountFeatureSelector<T>

For Beginners

Autocorrelation asks: “How similar is today’s value to yesterday’s, the day before, etc.?” Features with high autocorrelation have patterns that repeat or persist over time, making them potentially valuable for predicting future values.

How It Works

Autocorrelation measures how a feature correlates with itself at different time lags. Features with strong autocorrelation patterns often contain predictable structure that can be useful for time series forecasting.

Constructors

Constructor Summary
AutocorrelationFS(int nFeaturesToSelect, int maxLag, double minAutocorrelation, int[] columnIndices)

Properties

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

Methods

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