dragon.ml.seqmodel.crf

Class CollinsBasicTrainer

java.lang.Object

dragon.ml.seqmodel.crf.AbstractCRF

dragon.ml.seqmodel.crf.AbstractTrainer

dragon.ml.seqmodel.crf.CollinsBasicTrainer

All Implemented Interfaces:

Trainer

Direct Known Subclasses:

CollinsSegmentTrainer

public class CollinsBasicTrainer
extends AbstractTrainer

Collins training conditional random field

Copyright: Copyright (c) 2005

Company: IST, Drexel University

Version:

1.0

Author:

Davis Zhou

Field Summary

Fields

Modifier and Type	Field and Description
protected double	beta
protected int	topSolutions
protected boolean	useUpdated

Fields inherited from class dragon.ml.seqmodel.crf.AbstractTrainer

doScaling, maxIteration, xtol

Fields inherited from class dragon.ml.seqmodel.crf.AbstractCRF

edgeGen, featureGenerator, lambda, model

Constructor Summary

Constructors

Constructor and Description
CollinsBasicTrainer(ModelGraph model, FeatureGenerator featureGenerator)

Method Summary

Modifier and Type	Method and Description
protected Labeler	getLabeler()
protected int	getSegmentEnd(DataSequence dataSeq, int start)
protected double	getSequenceScore(DataSequence dataSeq, double[] grad)
protected boolean	isCorrect(DataSequence manual, DataSequence auto)
boolean	train(Dataset dataset) Trains the CRF model with labeled dataset
protected void	updateWeights(DataSequence dataSeq, int startPos, int endPos, double wt, double[] grad)

Methods inherited from class dragon.ml.seqmodel.crf.AbstractTrainer

genStateVector, genStateVectorLog, getMaxIteration, needScaling, setMaxIteration, setScalingOption

Methods inherited from class dragon.ml.seqmodel.crf.AbstractCRF

computeTransMatrix, computeTransMatrix, getFeatureGenerator, getModelGraph, getModelParameter, readModelParameter, saveModelParameter

Methods inherited from class java.lang.Object

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

Methods inherited from interface dragon.ml.seqmodel.crf.Trainer

getFeatureGenerator, getModelGraph, getModelParameter, saveModelParameter

Field Detail

topSolutions

protected int topSolutions

beta

protected double beta

useUpdated

protected boolean useUpdated

Constructor Detail

CollinsBasicTrainer

public CollinsBasicTrainer(ModelGraph model,
                           FeatureGenerator featureGenerator)

Method Detail

train

public boolean train(Dataset dataset)

Description copied from interface: Trainer

Trains the CRF model with labeled dataset

Parameters:

dataset - the dataset for training

Returns:

true if trained successfully

isCorrect

protected boolean isCorrect(DataSequence manual,
                            DataSequence auto)

updateWeights

protected void updateWeights(DataSequence dataSeq,
                             int startPos,
                             int endPos,
                             double wt,
                             double[] grad)

getSequenceScore

protected double getSequenceScore(DataSequence dataSeq,
                                  double[] grad)

getLabeler

protected Labeler getLabeler()

getSegmentEnd

protected int getSegmentEnd(DataSequence dataSeq,
                            int start)