org.apache.ignite.ml.recommendation

Class RecommendationTrainer

java.lang.Object

org.apache.ignite.ml.recommendation.RecommendationTrainer

public class RecommendationTrainer
extends Object

Trainer of the recommendation system.

Constructor Summary

Constructors

Constructor and Description
RecommendationTrainer()

Method Summary

Modifier and Type	Method and Description
<K,O extends Serializable,S extends Serializable> RecommendationModel<O,S>	fit(DatasetBuilder<K,? extends ObjectSubjectRatingTriplet<O,S>> datasetBuilder) Fits prediction model.
RecommendationModel<Serializable,Serializable>	fit(DatasetBuilder<Object,BinaryObject> datasetBuilder, String objFieldName, String subjFieldName, String ratingFieldName) Fits prediction model on a data storen in binary format.
RecommendationTrainer	withBatchSize(int batchSize) Set up batch size parameter.
RecommendationTrainer	withK(int k) Set up k parameter (number of rows/cols in matrices after factorization).
RecommendationTrainer	withLearningEnvironmentBuilder(LearningEnvironmentBuilder environmentBuilder) Set up learning environment builder.
RecommendationTrainer	withLearningRate(double learningRate) Set up learning rate parameter.
RecommendationTrainer	withMaxIterations(int maxIterations) Set up max iterations parameter.
RecommendationTrainer	withMinMdlImprovement(double minMdlImprovement) Set up minModelImprovement parameter (minimal improvement of the model to continue training).
RecommendationTrainer	withRegularizer(double regParam) Set up regularization parameter.
RecommendationTrainer	withTrainerEnvironment(LearningEnvironment trainerEnvironment) Set up trainer learning environment.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

RecommendationTrainer

public RecommendationTrainer()

Method Detail

fit

public RecommendationModel<Serializable,Serializable> fit(DatasetBuilder<Object,BinaryObject> datasetBuilder,
                                                          String objFieldName,
                                                          String subjFieldName,
                                                          String ratingFieldName)

Fits prediction model on a data storen in binary format.

Parameters:

datasetBuilder - Dataset builder.

objFieldName - Object field name.

subjFieldName - Subject field name.

ratingFieldName - Rating field name.

Returns:

Trained recommendation model.

fit

public <K,O extends Serializable,S extends Serializable> RecommendationModel<O,S> fit(DatasetBuilder<K,? extends ObjectSubjectRatingTriplet<O,S>> datasetBuilder)

Fits prediction model.

Type Parameters:

K - Type of a key in upstream data.

O - Type of an object.

S - Type of a subject.

Parameters:

datasetBuilder - Dataset builder.

Returns:

Trained recommendation model.

withLearningEnvironmentBuilder

public RecommendationTrainer withLearningEnvironmentBuilder(LearningEnvironmentBuilder environmentBuilder)

Set up learning environment builder.

Parameters:

environmentBuilder - Learning environment builder.

Returns:

This object.

withTrainerEnvironment

public RecommendationTrainer withTrainerEnvironment(LearningEnvironment trainerEnvironment)

Set up trainer learning environment.

Parameters:

trainerEnvironment - Trainer learning environment.

Returns:

This object.

withBatchSize

public RecommendationTrainer withBatchSize(int batchSize)

Set up batch size parameter.

Parameters:

batchSize - Batch size of stochastic gradient descent. The size of a dataset used on each step of SGD.

Returns:

This object.

withRegularizer

public RecommendationTrainer withRegularizer(double regParam)

Set up regularization parameter.

Parameters:

regParam - Regularization parameter.

Returns:

This object.

withLearningRate

public RecommendationTrainer withLearningRate(double learningRate)

Set up learning rate parameter.

Parameters:

learningRate - Learning rate.

Returns:

This object.

withMaxIterations

public RecommendationTrainer withMaxIterations(int maxIterations)

Set up max iterations parameter.

Parameters:

maxIterations - Max iterations.

Returns:

This object.

withMinMdlImprovement

public RecommendationTrainer withMinMdlImprovement(double minMdlImprovement)

Set up minModelImprovement parameter (minimal improvement of the model to continue training).

Parameters:

minMdlImprovement - Minimal improvement of the model to continue training.

Returns:

This object.

withK

public RecommendationTrainer withK(int k)

Set up k parameter (number of rows/cols in matrices after factorization).

Parameters:

k - Number of rows/cols in matrices after factorization

Returns:

This object.