iitb.CRF
Class CRF

java.lang.Object
  iitb.CRF.CRF

Direct Known Subclasses:

NestedCRF

public class CRF

extends java.lang.Object

CRF (conditional random fields) This class provides support for training and applying a conditional random field for sequence labeling problems.

Author:

Sunita Sarawagi

Field Summary

CrfParams

params

Constructor Summary

CRF(int numLabels, FeatureGenerator fgen, java.util.Properties configOptions)

CRF(int numLabels, FeatureGenerator fgen, java.lang.String arg)

Method Summary

void	apply(DataSequence dataSeq)
void	read(java.lang.String fileName) read the parameters of the CRF from a file
void	train(DataIter trainData)
void	write(java.lang.String fileName) write the trained parameters of the CRF to the file

Methods inherited from class java.lang.Object

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

Field Detail

params

public CrfParams params

Constructor Detail

CRF

public CRF(int numLabels,
           FeatureGenerator fgen,
           java.lang.String arg)

Parameters:

numLabels - is the number of distinct class labels or y-labels

fgen - is the class that is responsible for providing the features for a particular position on the sequence.

arg - is a string that can be used to control various parameters of the CRF, these are space separated name-value pairs described in

See Also:

CrfParams

CRF

public CRF(int numLabels,
           FeatureGenerator fgen,
           java.util.Properties configOptions)

Method Detail

write

public void write(java.lang.String fileName)
           throws java.io.IOException

write the trained parameters of the CRF to the file

Throws:

java.io.IOException

read

public void read(java.lang.String fileName)
          throws java.io.IOException

read the parameters of the CRF from a file

Throws:

java.io.IOException

train

public void train(DataIter trainData)

apply

public void apply(DataSequence dataSeq)