Class FMClassificationTrainer

java.lang.Object

org.tribuo.common.sgd.AbstractSGDTrainer<Label, Integer, FMClassificationModel, FMParameters>

org.tribuo.common.sgd.AbstractFMTrainer<Label, Integer, FMClassificationModel>

org.tribuo.classification.sgd.fm.FMClassificationTrainer

All Implemented Interfaces:

com.oracle.labs.mlrg.olcut.config.Configurable, com.oracle.labs.mlrg.olcut.provenance.Provenancable<org.tribuo.provenance.TrainerProvenance>, org.tribuo.Trainer<Label>, org.tribuo.WeightedExamples

public class FMClassificationTrainer
extends AbstractFMTrainer<Label, Integer, FMClassificationModel>

A trainer for a classification factorization machine using SGD.

See:

 Rendle, S.
 Factorization machines.
 2010 IEEE International Conference on Data Mining
 

Field Summary

Fields inherited from class org.tribuo.common.sgd.AbstractFMTrainer

factorizedDimSize, variance

Fields inherited from class org.tribuo.common.sgd.AbstractSGDTrainer

addBias, epochs, loggingInterval, minibatchSize, optimiser, rng, seed, shuffle

Fields inherited from interface org.tribuo.Trainer

DEFAULT_SEED, INCREMENT_INVOCATION_COUNT

Constructor Summary

Constructors

Constructor	Description
FMClassificationTrainer(LabelObjective objective, StochasticGradientOptimiser optimiser, int epochs, int loggingInterval, int minibatchSize, long seed, int factorizedDimSize, double variance)	Constructs an SGD trainer for a factorization machine.
FMClassificationTrainer(LabelObjective objective, StochasticGradientOptimiser optimiser, int epochs, int loggingInterval, long seed, int factorizedDimSize, double variance)	Constructs an SGD trainer for a factorization machine.
FMClassificationTrainer(LabelObjective objective, StochasticGradientOptimiser optimiser, int epochs, long seed, int factorizedDimSize, double variance)	Constructs an SGD trainer for a factorization machine.

Method Summary

Modifier and Type	Method	Description
protected FMClassificationModel	createModel(String name, org.tribuo.provenance.ModelProvenance provenance, org.tribuo.ImmutableFeatureMap featureMap, org.tribuo.ImmutableOutputInfo<Label> outputInfo, FMParameters parameters)
protected String	getModelClassName()
protected SGDObjective<Integer>	getObjective()
protected Integer	getTarget(org.tribuo.ImmutableOutputInfo<Label> outputInfo, Label output)
String	toString()

Methods inherited from class org.tribuo.common.sgd.AbstractFMTrainer

createParameters, getName, postConfig

Methods inherited from class org.tribuo.common.sgd.AbstractSGDTrainer

getInvocationCount, getProvenance, setInvocationCount, setShuffle, shuffleInPlace, train, train, train

Methods inherited from class java.lang.Object

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

Constructor Details

FMClassificationTrainer

public FMClassificationTrainer(LabelObjective objective,
 StochasticGradientOptimiser optimiser,
 int epochs,
 int loggingInterval,
 int minibatchSize,
 long seed,
 int factorizedDimSize,
 double variance)

Constructs an SGD trainer for a factorization machine.

Parameters:

objective - The objective function to optimise.

optimiser - The gradient optimiser to use.

epochs - The number of epochs (complete passes through the training data).

loggingInterval - Log the loss after this many iterations. If -1 don't log anything.

minibatchSize - The size of any minibatches.

seed - A seed for the random number generator, used to shuffle the examples before each epoch.

factorizedDimSize - Size of the factorized feature representation.

variance - The variance of the initializer.

FMClassificationTrainer

public FMClassificationTrainer(LabelObjective objective,
 StochasticGradientOptimiser optimiser,
 int epochs,
 int loggingInterval,
 long seed,
 int factorizedDimSize,
 double variance)

Constructs an SGD trainer for a factorization machine.

Sets the minibatch size to 1.

Parameters:

objective - The objective function to optimise.

optimiser - The gradient optimiser to use.

epochs - The number of epochs (complete passes through the training data).

loggingInterval - Log the loss after this many iterations. If -1 don't log anything.

seed - A seed for the random number generator, used to shuffle the examples before each epoch.

factorizedDimSize - Size of the factorized feature representation.

variance - The variance of the initializer.

FMClassificationTrainer

public FMClassificationTrainer(LabelObjective objective,
 StochasticGradientOptimiser optimiser,
 int epochs,
 long seed,
 int factorizedDimSize,
 double variance)

Constructs an SGD trainer for a factorization machine.

Sets the minibatch size to 1 and the logging interval to 1000.

Parameters:

objective - The objective function to optimise.

optimiser - The gradient optimiser to use.

epochs - The number of epochs (complete passes through the training data).

seed - A seed for the random number generator, used to shuffle the examples before each epoch.

factorizedDimSize - Size of the factorized feature representation.

variance - The variance of the initializer.

Method Details

getTarget

protected Integer getTarget(org.tribuo.ImmutableOutputInfo<Label> outputInfo,
 Label output)

Specified by:

getTarget in class AbstractSGDTrainer<Label, Integer, FMClassificationModel, FMParameters>

getObjective

protected SGDObjective<Integer> getObjective()

Specified by:

getObjective in class AbstractSGDTrainer<Label, Integer, FMClassificationModel, FMParameters>

createModel

protected FMClassificationModel createModel(String name,
 org.tribuo.provenance.ModelProvenance provenance,
 org.tribuo.ImmutableFeatureMap featureMap,
 org.tribuo.ImmutableOutputInfo<Label> outputInfo,
 FMParameters parameters)

Specified by:

createModel in class AbstractSGDTrainer<Label, Integer, FMClassificationModel, FMParameters>

getModelClassName

protected String getModelClassName()

Specified by:

getModelClassName in class AbstractSGDTrainer<Label, Integer, FMClassificationModel, FMParameters>

toString

public String toString()

Overrides:

toString in class Object