mark.nlp.classifiers
Class HierarchicalClassifier

java.lang.Object
  |
  +--mark.nlp.classifiers.HierarchicalClassifier

All Implemented Interfaces:

Classifier

Direct Known Subclasses:

ErrorCorrecting, HierarchyGenerating

public abstract class HierarchicalClassifier

extends java.lang.Object

implements Classifier

A base class for classifiers that classify in two steps, first through a root classifier and then through a child classifier.

Field Summary

protected mark.nlp.classifiers.Node	fC0
protected mark.nlp.classifiers.Node[]	fChildren
protected mark.nlp.classifiers.Node	fHC
protected int	fNumCats

Constructor Summary

HierarchicalClassifier()

Method Summary

int	classify(java.lang.Object instanceInformation) Classifies a single instance as returned from mark.nlp.classifiers.InstanceFactory.
protected abstract mark.nlp.classifiers.Node	genHierarchy(ObjectMap vocabulary, mark.nlp.classifiers.Node confusionClassifier, SimpleResults trainResults, FileCategorySet catSet, SparseBagInstance[] trainSparseBagInstances, java.lang.Object parameters)
protected static mark.nlp.classifiers.Node	genSubsetNode(FileCategorySet catSet, ObjectMap vocabulary, SparseBagInstance[] trainSparseBagInstances, java.lang.String featureOrdererName, java.lang.String featureOrdererParameter, int numFeatures, java.lang.String classifierName, java.lang.Object classifierParameter, int[] catsToDistinguish)
java.lang.String	instanceFactoryName() Returns the name of the instance factory that can generate instance information suitable for this classifier.
Results	test(java.lang.Object instanceListInformation) Tests the classifier on the given set of instances and returns results.
Results	test(java.lang.Object instanceListInformation, java.lang.String testIdentifier) Tests the classifier on the given set of instances and returns results.
void	train(java.lang.Object dataSetInformation, java.lang.Object instanceListInformation, java.lang.Object parameter) Trains the classifier.

Methods inherited from class java.lang.Object

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

Field Detail

fC0

protected mark.nlp.classifiers.Node fC0

fHC

protected mark.nlp.classifiers.Node fHC

fNumCats

protected int fNumCats

fChildren

protected mark.nlp.classifiers.Node[] fChildren

Constructor Detail

HierarchicalClassifier

public HierarchicalClassifier()

Method Detail

genSubsetNode

protected static mark.nlp.classifiers.Node genSubsetNode(FileCategorySet catSet,
                                                         ObjectMap vocabulary,
                                                         SparseBagInstance[] trainSparseBagInstances,
                                                         java.lang.String featureOrdererName,
                                                         java.lang.String featureOrdererParameter,
                                                         int numFeatures,
                                                         java.lang.String classifierName,
                                                         java.lang.Object classifierParameter,
                                                         int[] catsToDistinguish)
                                                  throws java.lang.Exception

java.lang.Exception

train

public void train(java.lang.Object dataSetInformation,
                  java.lang.Object instanceListInformation,
                  java.lang.Object parameter)
           throws java.lang.Exception

Description copied from interface: Classifier

Trains the classifier.

Specified by:

train in interface Classifier

Parameters:

dataSetInformation - the information returned from mark.nlp.classifiers.InstanceFactory common to all instances in the data set.

parameter - the classifier parameter which different classifiers interpret differently. A classifier may ignore this parameter.

java.lang.Exception

test

public Results test(java.lang.Object instanceListInformation)
             throws java.lang.Exception

Description copied from interface: Classifier

Tests the classifier on the given set of instances and returns results. The classifier must have been trained.

Specified by:

test in interface Classifier

Parameters:

instanceListInformation - the information returned from mark.nlp.classifiers.InstanceFactory specific to the list of instances on which we wish to test.

Returns:

claissification results.

java.lang.Exception

test

public Results test(java.lang.Object instanceListInformation,
                    java.lang.String testIdentifier)
             throws java.lang.Exception

Description copied from interface: Classifier

Tests the classifier on the given set of instances and returns results. The classifier must have been trained.

Specified by:

test in interface Classifier

Parameters:

instanceListInformation - the information returned from mark.nlp.classifiers.InstanceFactory specific to the list of instances on which we wish to test.

testIdentifier - A label included in the printed Results of the classifier.

Returns:

classification results.

java.lang.Exception

classify

public int classify(java.lang.Object instanceInformation)
             throws java.lang.Exception

Description copied from interface: Classifier

Classifies a single instance as returned from mark.nlp.classifiers.InstanceFactory. The classifier must have been trained.

Specified by:

classify in interface Classifier

Parameters:

instanceInformation - the information specific to the instance we wish to classify.

Returns:

the predicted category.

java.lang.Exception

instanceFactoryName

public java.lang.String instanceFactoryName()

Description copied from interface: Classifier

Returns the name of the instance factory that can generate instance information suitable for this classifier.

Specified by:

instanceFactoryName in interface Classifier

Returns:

the name.

genHierarchy

protected abstract mark.nlp.classifiers.Node genHierarchy(ObjectMap vocabulary,
                                                          mark.nlp.classifiers.Node confusionClassifier,
                                                          SimpleResults trainResults,
                                                          FileCategorySet catSet,
                                                          SparseBagInstance[] trainSparseBagInstances,
                                                          java.lang.Object parameters)
                                                   throws java.lang.Exception

java.lang.Exception