C# Class QACExperimenter.Approaches.SGD.OnlineSGDNonOverlappingLinearRegressionModel

Online training a non-overlapping linear regression model for the MNTB settings using SGD.

Afficher le fichier Open project: stewhir/recent-robust-qac Class Usage Examples

Méthodes publiques

Méthode	Description
MakePrediction ( double instanceIn ) : double	Publicly make prediction, as a query count (rather than scaled)
OnlineSGDNonOverlappingLinearRegressionModel ( int trainingFeatures, int maxNtbSize, int predictionHorizon ) : System	Derfault constructor
TrainModel ( FeaturePackage fp ) : void	Train the model with a FeaturePackage for a query

Private Methods

Méthode	Description
Predict ( double instance ) : double	Internal prediction method

Method Details

MakePrediction() public méthode

Publicly make prediction, as a query count (rather than scaled)

public MakePrediction ( double instanceIn ) : double
instanceIn	double
Résultat	double

OnlineSGDNonOverlappingLinearRegressionModel() public méthode

Derfault constructor

public OnlineSGDNonOverlappingLinearRegressionModel ( int trainingFeatures, int maxNtbSize, int predictionHorizon ) : System
trainingFeatures	int	Number of features to use
maxNtbSize	int
predictionHorizon	int
Résultat	System

TrainModel() public méthode

Train the model with a FeaturePackage for a query

public TrainModel ( FeaturePackage fp ) : void
fp	FeaturePackage
Résultat	void