Udostępnij za pośrednictwem


RegressionCatalog Klasa

Definicja

Klasa używana przez MLContext program do tworzenia wystąpień składników regresji, takich jak trenerzy i ewaluatorzy.

public sealed class RegressionCatalog : Microsoft.ML.TrainCatalogBase
type RegressionCatalog = class
    inherit TrainCatalogBase
Public NotInheritable Class RegressionCatalog
Inherits TrainCatalogBase
Dziedziczenie
RegressionCatalog

Właściwości

Trainers

Lista trenerów do wykonywania regresji.

Metody

CrossValidate(IDataView, IEstimator<ITransformer>, Int32, String, String, Nullable<Int32>)

Uruchom krzyżową walidację na numberOfFolds fałdach data, odpowiednio dopasowując estimatorelement i szanując samplingKeyColumnName , jeśli jest to podane. Następnie należy ocenić każdy podmodel pod kątem labelColumnName metryk i zwrócić je.

Evaluate(IDataView, String, String)

Ocenia wyniki danych regresji.

Metody rozszerzania

PermutationFeatureImportance(RegressionCatalog, ITransformer, IDataView, String, Boolean, Nullable<Int32>, Int32)

Ważność funkcji permutacji (PFI) dla regresji.

PermutationFeatureImportance<TModel>(RegressionCatalog, ISingleFeaturePredictionTransformer<TModel>, IDataView, String, Boolean, Nullable<Int32>, Int32)

Ważność funkcji permutacji (PFI) dla regresji.

Dotyczy