ISupervisedFeatureSelector<T>

Interfaces · AiDotNet.Interfaces

Defines a feature selector that needs the target values to decide which columns to keep.

Implements: IFeatureSelector<T>

Known implementations: ANCOVASelector<T>, ANOVAFTest<T>, ANOVASelector<T>, ANOVASelector<T>, ANOVA<T>, ARDSelector<T>, AdaptiveFeatureSelector<T>, AdjustedMutualInformation<T>, AlphaInvesting<T>, AntColonyFS<T>, AntColonyOptimization<T>, AntColonySelector<T>, AprioriBasedSelector<T>, ArtificialBeeColony<T>, AttentionBasedSelector<T>, AttentionFlowSelector<T>, BackwardElimination<T>, BaggingFeatureSelector<T>, BanzhafFeatureSelector<T>, BatAlgorithmFS<T>, BatAlgorithm<T>, BayesianFeatureSelector<T>, BayesianModelAveraging<T>, BayesianNetworkSelector<T>, BenjaminiHochberg<T>, BhattacharyyaDistanceSelector<T>, BidirectionalSearch<T>, BidirectionalSelection<T>, BoostingFeatureSelector<T>, BootstrapFS<T>, BootstrapSelection<T>, BootstrapSelector<T>, BootstrapStabilitySelector<T>, BorutaFS<T>, BorutaSelector<T>, Boruta<T>, Boruta<T>, BudgetConstrainedFS<T>, CFS<T>, CIFE<T>, CMIMSelector<T>, CMIM<T>, CMIM<T>, ChebyshevDistanceSelector<T>, ChiSquareTest<T>, ChiSquareText<T>, ChiSquare<T>, ChiSquare<T>, ChiSquaredSelection<T>, ClassWeightedFS<T>, ConditionalEntropySelector<T>, ConditionalEntropySelector<T>, ConstraintBasedSelector<T>, CorrelationBasedFS<T>, CorrelationClusterSelector<T>, CorrelationMIHybrid<T>, CosineSimilaritySelector<T>, CostSensitiveFS<T>, CrossValidatedSelector<T>, CuckooSearchFS<T> (+273 more)

For Beginners

These selectors ask “which columns help predict the answer?”, so they need the answers (the target) while fitting. Afterwards, transforming new data only needs the inputs.

How It Works

Filter methods that score each column against the target (mutual information, chi-squared, ANOVA F, ReliefF and similar) and wrapper methods that train a model cannot choose columns from the inputs alone, so they are fitted on the inputs together with the target.

Methods

Method Summary
Matrix<T> FitTransform(Matrix<T> data, Vector<T> target) Learns which columns to keep from the inputs and their target values, then returns the inputs with only those columns.
void Fit(Matrix<T> data, Vector<T> target) Learns which columns to keep from the inputs and their target values.