meka.classifiers.multilabel

Class LabelTransformationClassifier

java.lang.Object

weka.classifiers.AbstractClassifier

weka.classifiers.SingleClassifierEnhancer

meka.classifiers.multilabel.LabelTransformationClassifier

All Implemented Interfaces:

Serializable, Cloneable, MultiLabelClassifier, weka.classifiers.Classifier, weka.core.BatchPredictor, weka.core.CapabilitiesHandler, weka.core.CapabilitiesIgnorer, weka.core.CommandlineRunnable, weka.core.OptionHandler, weka.core.RevisionHandler

Direct Known Subclasses:

Maniac, MLCBMaD, PLST

public abstract class LabelTransformationClassifier
extends weka.classifiers.SingleClassifierEnhancer
implements MultiLabelClassifier

Abstract label transformation classifiers, all classes that transform the labels should inherit from this classifier. The general algorithm is a transformation of the labels into a new set of labels given a specific transformation method. A new multi-label classifier is then trained on the new set of latent labels. The prediction works the other way around, the multi-label classifier predicts the set of latent labels, and a specific transformation method transforms the predicted latent labels back into the original label dimension. Implementing classes need to implement three methods, first the transformation method used for the training process, second a transformation method to create a new instance as input for the prediction of the multi-label classifier, and third, a method transforming the predicted labels back into the original dimension.

Author:

Joerg Wicker (jw@joerg-wicker.org)

See Also:

Serialized Form

Field Summary

Fields inherited from class weka.classifiers.SingleClassifierEnhancer

m_Classifier

Fields inherited from class weka.classifiers.AbstractClassifier

BATCH_SIZE_DEFAULT, m_BatchSize, m_Debug, m_DoNotCheckCapabilities, m_numDecimalPlaces, NUM_DECIMAL_PLACES_DEFAULT

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	LabelTransformationClassifier() Default constructor using BR.

Method Summary

Modifier and Type	Method and Description
void	buildClassifier(weka.core.Instances D)
protected String	defaultClassifierString()
double[]	distributionForInstance(weka.core.Instance x)
protected weka.core.Instances	extractPart(weka.core.Instances inst, boolean labels) Returns a new set of instances either only with the labels (labels = true) or only the features (labels = false)
protected weka.classifiers.Classifier	getDefaultClassifier()
String	getRevision()
void	setClassifier(weka.classifiers.Classifier newClassifier)
void	testCapabilities(weka.core.Instances D) TestCapabilities.
abstract weka.core.Instance	transformInstance(weka.core.Instance x) Transforms the instance in the prediction process before given to the internal multi-label or multi-target classifier.
abstract weka.core.Instances	transformLabels(weka.core.Instances D) The method to transform the labels into another set of latent labels, typically a compression method is used, e.g., Boolean matrix decomposition in the case of MLC-BMaD, or matrix multiplication based on SVD for PLST.
abstract double[]	transformPredictionsBack(double[] y) Transforms the predictions of the internal classifier back to the original labels.

Methods inherited from class weka.classifiers.SingleClassifierEnhancer

classifierTipText, defaultClassifierOptions, getCapabilities, getClassifier, getClassifierSpec, getOptions, listOptions, postExecution, preExecution, setOptions

Methods inherited from class weka.classifiers.AbstractClassifier

batchSizeTipText, classifyInstance, debugTipText, distributionsForInstances, doNotCheckCapabilitiesTipText, forName, getBatchSize, getDebug, getDoNotCheckCapabilities, getNumDecimalPlaces, implementsMoreEfficientBatchPrediction, makeCopies, makeCopy, numDecimalPlacesTipText, run, runClassifier, setBatchSize, setDebug, setDoNotCheckCapabilities, setNumDecimalPlaces

Methods inherited from class java.lang.Object

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

Methods inherited from interface meka.classifiers.multilabel.MultiLabelClassifier

debugTipText, getDebug, getModel, setDebug

Methods inherited from interface weka.classifiers.Classifier

classifyInstance, getCapabilities

Methods inherited from interface weka.core.OptionHandler

getOptions, listOptions, setOptions

Constructor Detail

LabelTransformationClassifier

protected LabelTransformationClassifier()

Default constructor using BR.

Method Detail

transformLabels

public abstract weka.core.Instances transformLabels(weka.core.Instances D)
                                             throws Exception

The method to transform the labels into another set of latent labels, typically a compression method is used, e.g., Boolean matrix decomposition in the case of MLC-BMaD, or matrix multiplication based on SVD for PLST.

Parameters:

D - the instances to transform into new instances with transformed labels. The Instances consist of features and original labels.

Returns:

The resulting instances. Instances consist of features and transformed labels.

Throws:

Exception

transformInstance

public abstract weka.core.Instance transformInstance(weka.core.Instance x)
                                              throws Exception

Transforms the instance in the prediction process before given to the internal multi-label or multi-target classifier. The instance is passed having the original set of labels, these must be replaced with the transformed labels (attributes) so that the internla classifier can predict them.

Parameters:

x - The instance to transform. Consists of features and labels.

Returns:

The transformed instance. Consists of features and transformed labels.

Throws:

Exception

transformPredictionsBack

public abstract double[] transformPredictionsBack(double[] y)

Transforms the predictions of the internal classifier back to the original labels.

Parameters:

y - The predictions that should be transformed back. The array consists only of the predictions as they are returned from the internal classifier.

Returns:

The transformed predictions.

getDefaultClassifier

protected weka.classifiers.Classifier getDefaultClassifier()

defaultClassifierString

protected String defaultClassifierString()

Overrides:

defaultClassifierString in class weka.classifiers.SingleClassifierEnhancer

setClassifier

public void setClassifier(weka.classifiers.Classifier newClassifier)

Overrides:

setClassifier in class weka.classifiers.SingleClassifierEnhancer

extractPart

protected weka.core.Instances extractPart(weka.core.Instances inst,
                                          boolean labels)
                                   throws Exception

Returns a new set of instances either only with the labels (labels = true) or only the features (labels = false)

Parameters:

inst - The input instances.

labels - Return labels (true) or features (false)

Throws:

Exception

buildClassifier

public void buildClassifier(weka.core.Instances D)
                     throws Exception

Specified by:

buildClassifier in interface weka.classifiers.Classifier

Throws:

Exception

distributionForInstance

public double[] distributionForInstance(weka.core.Instance x)
                                 throws Exception

Specified by:

distributionForInstance in interface weka.classifiers.Classifier

Overrides:

distributionForInstance in class weka.classifiers.AbstractClassifier

Throws:

Exception

getRevision

public String getRevision()

Specified by:

getRevision in interface weka.core.RevisionHandler

Overrides:

getRevision in class weka.classifiers.AbstractClassifier

testCapabilities

public void testCapabilities(weka.core.Instances D)
                      throws Exception

TestCapabilities. Make sure the training data is suitable.

Parameters:

D - the data

Throws:

Exception