dragon.ml.seqmodel.crf

Interface Labeler

All Known Implementing Classes:

ViterbiBasicLabeler, ViterbiSegmentLabeler

public interface Labeler

Interface of CRF Labeler

Description:

Copyright: Copyright (c) 2005

Company: IST, Drexel University

Version:

1.0

Author:

Davis Zhou

Method Summary

Modifier and Type	Method and Description
double	getBestSolution(DataSequence dataSeq, int order) Gets the order-th solution.
FeatureGenerator	getFeatureGenerator() Gets the feature generator
ModelGraph	getModelGraph() Gets the model graph
double[]	getModelParameter() Gets the parameters of the CRF model
boolean	label(DataSequence dataSeq) Labels the given sequence
boolean	label(DataSequence dataSeq, double[] lambda) Labels the sequence using the given model
boolean	readModelParameter(String filename) Reads the parameters of the CRF model from a text file

Method Detail

getModelGraph

ModelGraph getModelGraph()

Gets the model graph

Returns:

the model graph

getFeatureGenerator

FeatureGenerator getFeatureGenerator()

Gets the feature generator

Returns:

the feature generator

getModelParameter

double[] getModelParameter()

Gets the parameters of the CRF model

Returns:

the parameters of the CRF model

readModelParameter

boolean readModelParameter(String filename)

Reads the parameters of the CRF model from a text file

Parameters:

filename - the name of the file containing the model parameters

Returns:

true if reading successfully

label

boolean label(DataSequence dataSeq)

Labels the given sequence

Parameters:

dataSeq - the sequence for labeling

Returns:

true if labeled successfully

label

boolean label(DataSequence dataSeq,
              double[] lambda)

Labels the sequence using the given model

Parameters:

dataSeq - the sequence for labeling

lambda - the parameters of the CRF model

Returns:

true if labeled successfully

getBestSolution

double getBestSolution(DataSequence dataSeq,
                       int order)

Gets the order-th solution. The labels will be set to the data sequence. To get the best solution, the order should be set to zero

Parameters:

dataSeq - the data seqeunce

order - the ranking of the solutions

Returns:

the score of the solution