dragon.ir.classification
Class AbstractClassifier

java.lang.Object
  |
  +--dragon.ir.classification.AbstractClassifier

All Implemented Interfaces:

Classifier

Direct Known Subclasses:

LibSVMClassifier, NBClassifier, SVMLightClassifier

public abstract class AbstractClassifier

extends java.lang.Object

implements Classifier

Basic function class for classifying

Copyright: Copyright (c) 2005

Company: IST, Drexel University

Version:

1.0

Author:

Davis Zhou

Field Summary

protected java.lang.String[]	arrLabel
protected int	classNum
protected SparseMatrix	doctermMatrix
protected FeatureSelector	featureSelector
protected IndexReader	indexReader
protected DocClassSet	validatingDocSet

Constructor Summary

AbstractClassifier()

AbstractClassifier(IndexReader indexReader)

AbstractClassifier(SparseMatrix doctermMatrix)

Method Summary

DocClassSet	classify(DocClass testingDocs) This method uses the trained model to classify the testing documents.
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.
DocClassSet	classify(DocClassSet trainingDocSet, DocClassSet validatingDocSet, DocClass testingDocs)
int	classify(IRDoc doc) Classify one particular document
java.lang.String	getClassLabel(int index) Gets the label of a given document category
SparseMatrix	getDocTermMatrix()
FeatureSelector	getFeatureSelector()
IndexReader	getIndexReader()
protected Row	getRow(int docIndex)
void	setFeatureSelector(FeatureSelector selector)
void	train(DocClassSet trainingDocSet, DocClassSet validatingDocSet) This method trains the classifier with the training document set and validating document set.
protected void	trainFeatureSelector(DocClassSet trainingSet)

Methods inherited from class java.lang.Object

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

Methods inherited from interface dragon.ir.classification.Classifier

classify, rank, saveModel, train

Field Detail

indexReader

protected IndexReader indexReader

doctermMatrix

protected SparseMatrix doctermMatrix

validatingDocSet

protected DocClassSet validatingDocSet

featureSelector

protected FeatureSelector featureSelector

arrLabel

protected java.lang.String[] arrLabel

classNum

protected int classNum

Constructor Detail

AbstractClassifier

public AbstractClassifier(IndexReader indexReader)

AbstractClassifier

public AbstractClassifier(SparseMatrix doctermMatrix)

AbstractClassifier

public AbstractClassifier()

Method Detail

getClassLabel

public java.lang.String getClassLabel(int index)

Description copied from interface: Classifier

Gets the label of a given document category

Specified by:

getClassLabel in interface Classifier

Parameters:

index - the index of the category

Returns:

the label of the category

getIndexReader

public IndexReader getIndexReader()

Specified by:

getIndexReader in interface Classifier

Returns:

the index reader the classifier iw working on

getDocTermMatrix

public SparseMatrix getDocTermMatrix()

getFeatureSelector

public FeatureSelector getFeatureSelector()

Specified by:

getFeatureSelector in interface Classifier

Returns:

the feature selector the current classifier is using

setFeatureSelector

public void setFeatureSelector(FeatureSelector selector)

Specified by:

setFeatureSelector in interface Classifier

Parameters:

selector - the feature selector for the classifier.

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

Parameters:

trainingDocSet - training document set

testingDocs - testing document set

Returns:

classified testing document set

classify

public DocClassSet classify(DocClassSet trainingDocSet,
                            DocClassSet validatingDocSet,
                            DocClass testingDocs)

Specified by:

classify in interface Classifier

Parameters:

trainingDocSet - the training document set

validatingDocSet - the validation document set, usually for avoiding the overfitting problem

testingDocs - the testing document set

Returns:

classified testing document set

train

public void train(DocClassSet trainingDocSet,
                  DocClassSet validatingDocSet)

Description copied from interface: Classifier

This method trains the classifier with the training document set and validating document set.

Specified by:

train in interface Classifier

Parameters:

trainingDocSet - training document set

validatingDocSet - validating document set

classify

public DocClassSet classify(DocClass testingDocs)

Description copied from interface: Classifier

This method uses the trained model to classify the testing documents. The train method should be called before calling this method.

Specified by:

classify in interface Classifier

Parameters:

testingDocs - testing document set

Returns:

classified testing document set

classify

public int classify(IRDoc doc)

Description copied from interface: Classifier

Classify one particular document

Specified by:

classify in interface Classifier

Parameters:

doc - 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.

getRow

protected Row getRow(int docIndex)

trainFeatureSelector

protected void trainFeatureSelector(DocClassSet trainingSet)