eus.ixa.ixa.pipe.nerc.train

Class AbstractTrainer

java.lang.Object

eus.ixa.ixa.pipe.nerc.train.AbstractTrainer

All Implemented Interfaces:

Trainer

Direct Known Subclasses:

FixedTrainer

public abstract class AbstractTrainer
extends Object
implements Trainer

Abstract class for common training functionalities. Every other trainer class needs to extend this class.

Version:

2015-02-25

Author:

ragerri

Constructor Summary

Constructors

Constructor and Description
AbstractTrainer(opennlp.tools.util.TrainingParameters params) Construct a trainer with training and test data, and with options for language, beamsize for decoding, sequence codec and corpus format (conll or opennlp).

Method Summary

Modifier and Type	Method and Description
int	getBeamSize()
String	getLanguage() Get the language.
opennlp.tools.namefind.TokenNameFinderFactory	getNameClassifierFactory() Get the features which are implemented in each of the trainers extending this class.
static opennlp.tools.util.ObjectStream<opennlp.tools.namefind.NameSample>	getNameStream(String inputData, String clearFeatures, String aCorpusFormat) Getting the stream with the right corpus format.
String	getSequenceCodec() Get the Sequence codec.
void	setLanguage(String aLang) Set the language.
opennlp.tools.namefind.TokenNameFinderFactory	setNameClassifierFactory(opennlp.tools.namefind.TokenNameFinderFactory tokenNameFinderFactory)
void	setSequenceCodec(String aSeqCodec) Set the sequence codec.
opennlp.tools.namefind.TokenNameFinderModel	train(opennlp.tools.util.TrainingParameters params) Generate TokenNameFinderModel models.

Methods inherited from class java.lang.Object

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

Constructor Detail

AbstractTrainer

public AbstractTrainer(opennlp.tools.util.TrainingParameters params)
                throws IOException

Construct a trainer with training and test data, and with options for language, beamsize for decoding, sequence codec and corpus format (conll or opennlp).

Parameters:

params - the training parameters

Throws:

IOException - io exception

Method Detail

train

public final opennlp.tools.namefind.TokenNameFinderModel train(opennlp.tools.util.TrainingParameters params)

Description copied from interface: Trainer

Generate TokenNameFinderModel models.

Specified by:

train in interface Trainer

Parameters:

params - the training parameters file

Returns:

the model

getNameStream

public static opennlp.tools.util.ObjectStream<opennlp.tools.namefind.NameSample> getNameStream(String inputData,
                                                                                               String clearFeatures,
                                                                                               String aCorpusFormat)
                                                                                        throws IOException

Getting the stream with the right corpus format.

Parameters:

inputData - the input data

clearFeatures - clear the features

aCorpusFormat - the corpus format

Returns:

the stream from the several corpus formats

Throws:

IOException - the io exception

getNameClassifierFactory

public final opennlp.tools.namefind.TokenNameFinderFactory getNameClassifierFactory()

Get the features which are implemented in each of the trainers extending this class.

Returns:

the features

setNameClassifierFactory

public final opennlp.tools.namefind.TokenNameFinderFactory setNameClassifierFactory(opennlp.tools.namefind.TokenNameFinderFactory tokenNameFinderFactory)

getLanguage

public final String getLanguage()

Get the language.

Returns:

the language

setLanguage

public final void setLanguage(String aLang)

Set the language.

Parameters:

aLang - the language

getSequenceCodec

public final String getSequenceCodec()

Get the Sequence codec.

Returns:

the sequence codec

setSequenceCodec

public final void setSequenceCodec(String aSeqCodec)

Set the sequence codec.

Parameters:

aSeqCodec - the sequence codec to be set

getBeamSize

public final int getBeamSize()