dragon.ir.classification

Class NigamActiveLearning

java.lang.Object

dragon.ir.classification.AbstractClassifier

dragon.ir.classification.NBClassifier

dragon.ir.classification.NigamActiveLearning

All Implemented Interfaces:

Classifier

public class NigamActiveLearning
extends NBClassifier

Nigam Active Learning which can utilize unlabeled documents during text classification

Please refer to the following paper for details:
Nigam, K., McCallum, A., Thrun, S., Mitchell, T. Text Classification from Labeled and Unlabeled Documents using EM, Machine Learning, Volume 39, Issue 2-3 (May-June 2000), pp103-134

Copyright: Copyright (c) 2005

Company: IST, Drexel University

Version:

1.0

Author:

Davis Zhou

Field Summary

Fields inherited from class dragon.ir.classification.NBClassifier

classPrior, lastClassProb, model

Fields inherited from class dragon.ir.classification.AbstractClassifier

arrLabel, classNum, doctermMatrix, featureSelector, indexReader, validatingDocSet

Constructor Summary

Constructors

Constructor and Description
NigamActiveLearning(IndexReader indexReader, double unlabeledRate)
NigamActiveLearning(String modelFile)

Method Summary

Modifier and Type	Method and Description
DocClassSet	classify(DocClassSet trainingDocSet, DocClass testingDocs) This method trains the classifier with the training document set and then using the trained model to classify the testing documents.
int	classify(IRDoc curDoc) Classify one particular document
void	setUnlabeledData(DocClass docSet)
void	setUnlabeledData(IndexReader newIndexReader, DocClass docSet)
void	train(DocClassSet trainingDocSet) This method trains the classifier with the training document set.

Methods inherited from class dragon.ir.classification.NBClassifier

classify, getClassPrior, rank, saveModel

Methods inherited from class dragon.ir.classification.AbstractClassifier

classify, classify, getClassLabel, getDocTermMatrix, getFeatureSelector, getIndexReader, getRow, setFeatureSelector, train, trainFeatureSelector

Methods inherited from class java.lang.Object

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

Constructor Detail

NigamActiveLearning

public NigamActiveLearning(String modelFile)

NigamActiveLearning

public NigamActiveLearning(IndexReader indexReader,
                           double unlabeledRate)

Method Detail

setUnlabeledData

public void setUnlabeledData(IndexReader newIndexReader,
                             DocClass docSet)

setUnlabeledData

public void setUnlabeledData(DocClass docSet)

classify

public DocClassSet classify(DocClassSet trainingDocSet,
                            DocClass testingDocs)

Description copied from interface: Classifier

This method trains the classifier with the training document set and then using the trained model to classify the testing documents.

Specified by:

classify in interface Classifier

Overrides:

classify in class AbstractClassifier

Parameters:

trainingDocSet - training document set

testingDocs - testing document set

Returns:

classified testing document set

train

public void train(DocClassSet trainingDocSet)

Description copied from interface: Classifier

This method trains the classifier with the training document set.

Specified by:

train in interface Classifier

Overrides:

train in class NBClassifier

Parameters:

trainingDocSet - training document set

classify

public int classify(IRDoc curDoc)

Description copied from interface: Classifier

Classify one particular document

Specified by:

classify in interface Classifier

Overrides:

classify in class NBClassifier

Parameters:

curDoc - the index of the document is stored in the IRDoc object

Returns:

the index of the category of this document. The index starts from zero.