Package org.tribuo.common.sgd

Class AbstractFMTrainer<T extends Output<T>,U,V extends AbstractFMModel<T>>

java.lang.Object

org.tribuo.common.sgd.AbstractSGDTrainer<T,U,V,FMParameters>

org.tribuo.common.sgd.AbstractFMTrainer<T,U,V>

All Implemented Interfaces:

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

Direct Known Subclasses:

FMClassificationTrainer, FMMultiLabelTrainer, FMRegressionTrainer

public abstract class AbstractFMTrainer<T extends Output<T>,U,V extends AbstractFMModel<T>>
extends AbstractSGDTrainer<T,U,V,FMParameters>

A trainer for a quadratic factorization machine model which uses SGD.

It's an AbstractSGDTrainer operating on FMParameters.

See:

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

Field Summary

Fields

Modifier and Type	Field	Description
protected int	factorizedDimSize
protected double	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

Modifier	Constructor	Description
protected	AbstractFMTrainer()	For olcut.
protected	AbstractFMTrainer(StochasticGradientOptimiser optimiser, int epochs, int loggingInterval, int minibatchSize, long seed, int factorizedDimSize, double variance)	Constructs an SGD trainer for a factorization machine.

Method Summary

Modifier and Type	Method	Description
protected FMParameters	createParameters(int numFeatures, int numOutputs, SplittableRandom localRNG)	Constructs the trainable parameters object, in this case a FMParameters containing a weight matrix for the feature weights and a series of weight matrices for the factorized feature representation.
protected String	getName()	Returns the default model name.
void	postConfig()	Used by the OLCUT configuration system, and should not be called by external code.

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

createModel, getInvocationCount, getModelClassName, getObjective, getProvenance, getTarget, setInvocationCount, setShuffle, shuffleInPlace, train, train, train

Methods inherited from class java.lang.Object

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

Field Details

factorizedDimSize

@Config(mandatory=true,
        description="The size of the factorized feature representation.")
protected int factorizedDimSize

variance

@Config(mandatory=true,
        description="The variance of the initializer.")
protected double variance

Constructor Details

AbstractFMTrainer

protected AbstractFMTrainer(StochasticGradientOptimiser optimiser,
 int epochs,
 int loggingInterval,
 int minibatchSize,
 long seed,
 int factorizedDimSize,
 double variance)

Constructs an SGD trainer for a factorization machine.

Parameters:

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.

AbstractFMTrainer

protected AbstractFMTrainer()

For olcut.

Method Details

postConfig

public void postConfig()

Description copied from class: AbstractSGDTrainer

Used by the OLCUT configuration system, and should not be called by external code.

Specified by:

postConfig in interface com.oracle.labs.mlrg.olcut.config.Configurable

Overrides:

postConfig in class AbstractSGDTrainer<T extends Output<T>,U,V extends AbstractFMModel<T>,FMParameters>

getName

protected String getName()

Returns the default model name.

Specified by:

getName in class AbstractSGDTrainer<T extends Output<T>,U,V extends AbstractFMModel<T>,FMParameters>

Returns:

The default model name.

createParameters

protected FMParameters createParameters(int numFeatures,
 int numOutputs,
 SplittableRandom localRNG)

Constructs the trainable parameters object, in this case a FMParameters containing a weight matrix for the feature weights and a series of weight matrices for the factorized feature representation.

Specified by:

createParameters in class AbstractSGDTrainer<T extends Output<T>,U,V extends AbstractFMModel<T>,FMParameters>

Parameters:

numFeatures - The number of input features.

numOutputs - The number of output dimensions.

localRNG - The RNG to use for parameter initialisation.

Returns:

The trainable parameters.