opennlp.tools.ml.maxent

Class GISTrainer

java.lang.Object

opennlp.tools.ml.AbstractTrainer

opennlp.tools.ml.AbstractEventTrainer

opennlp.tools.ml.maxent.GISTrainer

All Implemented Interfaces:

EventTrainer

public class GISTrainer
extends AbstractEventTrainer

An implementation of Generalized Iterative Scaling. The reference paper for this implementation was Adwait Ratnaparkhi's tech report at the University of Pennsylvania's Institute for Research in Cognitive Science, and is available at ftp://ftp.cis.upenn.edu/pub/ircs/tr/97-08.ps.Z.

The slack parameter used in the above implementation has been removed by default from the computation and a method for updating with Gaussian smoothing has been added per Investigating GIS and Smoothing for Maximum Entropy Taggers, Clark and Curran (2002). http://acl.ldc.upenn.edu/E/E03/E03-1071.pdf The slack parameter can be used by setting useSlackParameter to true. Gaussian smoothing can be used by setting useGaussianSmoothing to true.

A prior can be used to train models which converge to the distribution which minimizes the relative entropy between the distribution specified by the empirical constraints of the training data and the specified prior. By default, the uniform distribution is used as the prior.

Field Summary

Fields

Modifier and Type	Field and Description
static double	LOG_LIKELIHOOD_THRESHOLD_DEFAULT
static java.lang.String	LOG_LIKELIHOOD_THRESHOLD_PARAM
static java.lang.String	MAXENT_VALUE
static java.lang.String	OLD_LL_THRESHOLD_PARAM Deprecated.

Fields inherited from class opennlp.tools.ml.AbstractEventTrainer

DATA_INDEXER_ONE_PASS_REAL_VALUE, DATA_INDEXER_ONE_PASS_VALUE, DATA_INDEXER_PARAM, DATA_INDEXER_TWO_PASS_VALUE

Fields inherited from class opennlp.tools.ml.AbstractTrainer

ALGORITHM_PARAM, CUTOFF_DEFAULT, CUTOFF_PARAM, ITERATIONS_DEFAULT, ITERATIONS_PARAM, TRAINER_TYPE_PARAM, VERBOSE_DEFAULT, VERBOSE_PARAM

Fields inherited from interface opennlp.tools.ml.EventTrainer

EVENT_VALUE

Constructor Summary

Constructors

Constructor and Description
GISTrainer() Creates a new GISTrainer instance which does not print progress messages about training to STDOUT.

Method Summary

Modifier and Type	Method and Description
MaxentModel	doTrain(DataIndexer indexer)
void	init(TrainingParameters trainingParameters, java.util.Map<java.lang.String,java.lang.String> reportMap)
boolean	isSortAndMerge()
void	setGaussianSigma(double sigmaValue) Sets whether this trainer will use smoothing while training the model.
void	setSmoothing(boolean smooth) Sets whether this trainer will use smoothing while training the model.
void	setSmoothingObservation(double timesSeen) Sets whether this trainer will use smoothing while training the model.
GISModel	trainModel(int iterations, DataIndexer di) Train a model using the GIS algorithm.
GISModel	trainModel(int iterations, DataIndexer di, int threads) Train a model using the GIS algorithm.
GISModel	trainModel(int iterations, DataIndexer di, Prior modelPrior, int threads) Train a model using the GIS algorithm.
GISModel	trainModel(ObjectStream<Event> eventStream) Train a model using the GIS algorithm, assuming 100 iterations and no cutoff.
GISModel	trainModel(ObjectStream<Event> eventStream, int iterations, int cutoff) Trains a GIS model on the event in the specified event stream, using the specified number of iterations and the specified count cutoff.

Methods inherited from class opennlp.tools.ml.AbstractEventTrainer

getDataIndexer, isValid, train, train, validate

Methods inherited from class opennlp.tools.ml.AbstractTrainer

getAlgorithm, getCutoff, getIterations, init

Methods inherited from class java.lang.Object

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

Methods inherited from interface opennlp.tools.ml.EventTrainer

init

Field Detail

OLD_LL_THRESHOLD_PARAM

@Deprecated
public static final java.lang.String OLD_LL_THRESHOLD_PARAM

Deprecated.

See Also:

Constant Field Values

LOG_LIKELIHOOD_THRESHOLD_PARAM

public static final java.lang.String LOG_LIKELIHOOD_THRESHOLD_PARAM

See Also:

Constant Field Values

LOG_LIKELIHOOD_THRESHOLD_DEFAULT

public static final double LOG_LIKELIHOOD_THRESHOLD_DEFAULT

See Also:

Constant Field Values

MAXENT_VALUE

public static final java.lang.String MAXENT_VALUE

See Also:

Constant Field Values

Constructor Detail

GISTrainer

public GISTrainer()

Creates a new GISTrainer instance which does not print progress messages about training to STDOUT.

Method Detail

isSortAndMerge

public boolean isSortAndMerge()

Specified by:

isSortAndMerge in class AbstractEventTrainer

init

public void init(TrainingParameters trainingParameters,
                 java.util.Map<java.lang.String,java.lang.String> reportMap)

Specified by:

init in interface EventTrainer

Overrides:

init in class AbstractTrainer

doTrain

public MaxentModel doTrain(DataIndexer indexer)
                    throws java.io.IOException

Specified by:

doTrain in class AbstractEventTrainer

Throws:

java.io.IOException

setSmoothing

public void setSmoothing(boolean smooth)

Sets whether this trainer will use smoothing while training the model. This can improve model accuracy, though training will potentially take longer and use more memory. Model size will also be larger.

Parameters:

smooth - true if smoothing is desired, false if not

setSmoothingObservation

public void setSmoothingObservation(double timesSeen)

Sets whether this trainer will use smoothing while training the model. This can improve model accuracy, though training will potentially take longer and use more memory. Model size will also be larger.

Parameters:

timesSeen - the "number" of times we want the trainer to imagine it saw a feature that it actually didn't see

setGaussianSigma

public void setGaussianSigma(double sigmaValue)

Sets whether this trainer will use smoothing while training the model. This can improve model accuracy, though training will potentially take longer and use more memory. Model size will also be larger.

trainModel

public GISModel trainModel(ObjectStream<Event> eventStream)
                    throws java.io.IOException

Train a model using the GIS algorithm, assuming 100 iterations and no cutoff.

Parameters:

eventStream - The EventStream holding the data on which this model will be trained.

Returns:

The newly trained model, which can be used immediately or saved to disk using an opennlp.tools.ml.maxent.io.GISModelWriter object.

Throws:

java.io.IOException

trainModel

public GISModel trainModel(ObjectStream<Event> eventStream,
                           int iterations,
                           int cutoff)
                    throws java.io.IOException

Trains a GIS model on the event in the specified event stream, using the specified number of iterations and the specified count cutoff.

Parameters:

eventStream - A stream of all events.

iterations - The number of iterations to use for GIS.

cutoff - The number of times a feature must occur to be included.

Returns:

A GIS model trained with specified

Throws:

java.io.IOException

trainModel

public GISModel trainModel(int iterations,
                           DataIndexer di)

Train a model using the GIS algorithm.

Parameters:

iterations - The number of GIS iterations to perform.

di - The data indexer used to compress events in memory.

Returns:

The newly trained model, which can be used immediately or saved to disk using an opennlp.tools.ml.maxent.io.GISModelWriter object.

trainModel

public GISModel trainModel(int iterations,
                           DataIndexer di,
                           int threads)

Train a model using the GIS algorithm.

Parameters:

iterations - The number of GIS iterations to perform.

di - The data indexer used to compress events in memory.

threads -

Returns:

The newly trained model, which can be used immediately or saved to disk using an opennlp.tools.ml.maxent.io.GISModelWriter object.

trainModel

public GISModel trainModel(int iterations,
                           DataIndexer di,
                           Prior modelPrior,
                           int threads)

Train a model using the GIS algorithm.

Parameters:

iterations - The number of GIS iterations to perform.

di - The data indexer used to compress events in memory.

modelPrior - The prior distribution used to train this model.

Returns:

The newly trained model, which can be used immediately or saved to disk using an opennlp.tools.ml.maxent.io.GISModelWriter object.