weka.classifiers.meta
Class Decorate

java.lang.Object
  weka.classifiers.Classifier
      weka.classifiers.meta.Decorate

All Implemented Interfaces:

java.lang.Cloneable, OptionHandler, java.io.Serializable

public class Decorate

extends Classifier

implements OptionHandler

DECORATE is a meta-learner for building diverse ensembles of classifiers by using specially constructed artificial training examples. Comprehensive experiments have demonstrated that this technique is consistently more accurate than the base classifier, Bagging and Random Forests. Decorate also obtains higher accuracy than Boosting on small training sets, and achieves comparable performance on larger training sets. For more details see:

Prem Melville and Raymond J. Mooney. Constructing diverse classifier ensembles using artificial training examples. Proceedings of the Seventeeth International Joint Conference on Artificial Intelligence 2003.

Prem Melville and Raymond J. Mooney. Creating diversity in ensembles using artificial data. Submitted.

Valid options are:

-D
Turn on debugging output.

-W classname
Specify the full class name of a weak classifier as the basis for Decorate (default weka.classifiers.trees.J48()).

-I num
Specify the desired size of the committee (default 15).

-M iterations
Set the maximum number of Decorate iterations (default 50).

-S seed
Seed for random number generator. (default 0).

-R factor
Factor that determines number of artificial examples to generate.

Options after -- are passed to the designated classifier.

Version:

$Revision: 1.2 $

Author:

Prem Melville (melville@cs.utexas.edu)

See Also:

Serialized Form

Field Summary

protected double	m_ArtSize Amount of artificial/random instances to use - specified as a fraction of the training data size.
protected java.util.Vector	m_AttributeStats Attribute statistics - used for generating artificial examples.
protected Classifier	m_Classifier The model base classifier to use.
protected java.util.Vector	m_Committee Vector of classifiers that make up the committee/ensemble.
protected boolean	m_Debug Set to true to get debugging output.
protected int	m_DesiredSize The desired ensemble size.
protected int	m_NumIterations The maximum number of Decorate iterations to run.
protected java.util.Random	m_Random The random number generator.
protected int	m_Seed The seed for random number generation.

Constructor Summary

Decorate()

Method Summary

protected void	addInstances(Instances data, Instances newData) Add new instances to the given set of instances.
java.lang.String	artificialSizeTipText() Returns the tip text for this property
void	buildClassifier(Instances data) Build Decorate classifier
java.lang.String	classifierTipText() Returns the tip text for this property
protected double	computeError(Instances data) Computes the error in classification on the given data.
protected void	computeStats(Instances data) Compute and store statistics required for generating artificial data.
java.lang.String	desiredSizeTipText() Returns the tip text for this property
double[]	distributionForInstance(Instance instance) Calculates the class membership probabilities for the given test instance.
protected Instances	generateArtificialData(int artSize, Instances data) Generate artificial training examples.
double	getArtificialSize() Factor that determines number of artificial examples to generate.
Classifier	getClassifier() Get the classifier used as the base classifier
boolean	getDebug() Get whether debugging is turned on
int	getDesiredSize() Gets the desired size of the committee.
int	getNumIterations() Gets the max number of Decorate iterations to run.
java.lang.String[]	getOptions() Gets the current settings of the Classifier.
int	getSeed() Gets the seed for the random number generator.
java.lang.String	globalInfo() Returns a string describing classifier
protected int	inverseLabel(double[] probs) Select class label such that the probability of selection is inversely proportional to the ensemble's predictions.
protected void	labelData(Instances artData) Labels the artificially generated data.
java.util.Enumeration	listOptions() Returns an enumeration describing the available options
static void	main(java.lang.String[] argv) Main method for testing this class.
java.lang.String	numIterationsTipText() Returns the tip text for this property
protected void	removeInstances(Instances data, int numRemove) Removes a specified number of instances from the given set of instances.
java.lang.String	seedTipText() Returns the tip text for this property
protected int	selectIndexProbabilistically(double[] cdf) Given cumulative probabilities select a nominal attribute value index
void	setArtificialSize(double newArtSize) Sets factor that determines number of artificial examples to generate.
void	setClassifier(Classifier newClassifier) Set the base classifier for Decorate.
void	setDebug(boolean debug) Set debugging mode
void	setDesiredSize(int newDesiredSize) Sets the desired size of the committee.
void	setNumIterations(int numIterations) Sets the max number of Decorate iterations to run.
void	setOptions(java.lang.String[] options) Parses a given list of options.
void	setSeed(int seed) Set the seed for random number generator.
java.lang.String	toString() Returns description of the Decorate classifier.

Methods inherited from class weka.classifiers.Classifier

classifyInstance, debugTipText, forName, makeCopies

Methods inherited from class java.lang.Object

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

Field Detail

m_Debug

protected boolean m_Debug

Set to true to get debugging output.

m_Classifier

protected Classifier m_Classifier

The model base classifier to use.

m_Committee

protected java.util.Vector m_Committee

Vector of classifiers that make up the committee/ensemble.

m_DesiredSize

protected int m_DesiredSize

The desired ensemble size.

m_NumIterations

protected int m_NumIterations

The maximum number of Decorate iterations to run.

m_Seed

protected int m_Seed

The seed for random number generation.

m_ArtSize

protected double m_ArtSize

Amount of artificial/random instances to use - specified as a fraction of the training data size.

m_Random

protected java.util.Random m_Random

The random number generator.

m_AttributeStats

protected java.util.Vector m_AttributeStats

Attribute statistics - used for generating artificial examples.

Constructor Detail

Decorate

public Decorate()

Method Detail

listOptions

public java.util.Enumeration listOptions()

Returns an enumeration describing the available options

Specified by:

listOptions in interface OptionHandler

Overrides:

listOptions in class Classifier

Returns:

an enumeration of all the available options

setOptions

public void setOptions(java.lang.String[] options)
                throws java.lang.Exception

Parses a given list of options. Valid options are:

-D
Turn on debugging output.

-W classname
Specify the full class name of a weak classifier as the basis for Decorate (required).

-I num
Specify the desired size of the committee (default 15).

-M iterations
Set the maximum number of Decorate iterations (default 50).

-S seed
Seed for random number generator. (default 0).

-R factor
Factor that determines number of artificial examples to generate.

Options after -- are passed to the designated classifier.

Specified by:

setOptions in interface OptionHandler

Overrides:

setOptions in class Classifier

Parameters:

options - the list of options as an array of strings

Throws:

java.lang.Exception - if an option is not supported

getOptions

public java.lang.String[] getOptions()

Gets the current settings of the Classifier.

Specified by:

getOptions in interface OptionHandler

Overrides:

getOptions in class Classifier

Returns:

an array of strings suitable for passing to setOptions

desiredSizeTipText

public java.lang.String desiredSizeTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui

numIterationsTipText

public java.lang.String numIterationsTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui

artificialSizeTipText

public java.lang.String artificialSizeTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui

seedTipText

public java.lang.String seedTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui

classifierTipText

public java.lang.String classifierTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui b

globalInfo

public java.lang.String globalInfo()

Returns a string describing classifier

Returns:

a description suitable for displaying in the explorer/experimenter gui

setDebug

public void setDebug(boolean debug)

Set debugging mode

Overrides:

setDebug in class Classifier

Parameters:

debug - true if debug output should be printed

getDebug

public boolean getDebug()

Get whether debugging is turned on

Overrides:

getDebug in class Classifier

Returns:

true if debugging output is on

setClassifier

public void setClassifier(Classifier newClassifier)

Set the base classifier for Decorate.

Parameters:

newClassifier - the Classifier to use.

getClassifier

public Classifier getClassifier()

Get the classifier used as the base classifier

Returns:

the classifier used as the classifier

getArtificialSize

public double getArtificialSize()

Factor that determines number of artificial examples to generate.

Returns:

factor that determines number of artificial examples to generate

setArtificialSize

public void setArtificialSize(double newArtSize)

Sets factor that determines number of artificial examples to generate.

getDesiredSize

public int getDesiredSize()

Gets the desired size of the committee.

Returns:

the desired size of the committee

setDesiredSize

public void setDesiredSize(int newDesiredSize)

Sets the desired size of the committee.

Parameters:

newDesiredSize - the desired size of the committee

setNumIterations

public void setNumIterations(int numIterations)

Sets the max number of Decorate iterations to run.

Parameters:

numIterations - max number of Decorate iterations to run

getNumIterations

public int getNumIterations()

Gets the max number of Decorate iterations to run.

Returns:

the max number of Decorate iterations to run

setSeed

public void setSeed(int seed)

Set the seed for random number generator.

Parameters:

seed - the random number seed

getSeed

public int getSeed()

Gets the seed for the random number generator.

Returns:

the seed for the random number generator

buildClassifier

public void buildClassifier(Instances data)
                     throws java.lang.Exception

Build Decorate classifier

Specified by:

buildClassifier in class Classifier

Parameters:

data - the training data to be used for generating the classifier

Throws:

java.lang.Exception - if the classifier could not be built successfully

computeStats

protected void computeStats(Instances data)
                     throws java.lang.Exception

Compute and store statistics required for generating artificial data.

Parameters:

data - training instances

Throws:

java.lang.Exception - if statistics could not be calculated successfully

generateArtificialData

protected Instances generateArtificialData(int artSize,
                                           Instances data)

Generate artificial training examples.

Parameters:

artSize - size of examples set to create

data - training data

Returns:

the set of unlabeled artificial examples

labelData

protected void labelData(Instances artData)
                  throws java.lang.Exception

Labels the artificially generated data.

Parameters:

artData - the artificially generated instances

Throws:

java.lang.Exception - if instances cannot be labeled successfully

inverseLabel

protected int inverseLabel(double[] probs)
                    throws java.lang.Exception

Select class label such that the probability of selection is inversely proportional to the ensemble's predictions.

Parameters:

probs - class membership probabilities of instance

Returns:

index of class label selected

Throws:

java.lang.Exception - if instances cannot be labeled successfully

selectIndexProbabilistically

protected int selectIndexProbabilistically(double[] cdf)

Given cumulative probabilities select a nominal attribute value index

Parameters:

cdf - array of cumulative probabilities

Returns:

index of attribute selected based on the probability distribution

removeInstances

protected void removeInstances(Instances data,
                               int numRemove)

Removes a specified number of instances from the given set of instances.

Parameters:

data - given instances

numRemove - number of instances to delete from the given instances

addInstances

protected void addInstances(Instances data,
                            Instances newData)

Add new instances to the given set of instances.

Parameters:

data - given instances

newData - set of instances to add to given instances

computeError

protected double computeError(Instances data)
                       throws java.lang.Exception

Computes the error in classification on the given data.

Parameters:

data - the instances to be classified

Returns:

classification error

Throws:

java.lang.Exception - if error can not be computed successfully

distributionForInstance

public double[] distributionForInstance(Instance instance)
                                 throws java.lang.Exception

Calculates the class membership probabilities for the given test instance.

Overrides:

distributionForInstance in class Classifier

Parameters:

instance - the instance to be classified

Returns:

predicted class probability distribution

Throws:

java.lang.Exception - if distribution can't be computed successfully

toString

public java.lang.String toString()

Returns description of the Decorate classifier.

Returns:

description of the Decorate classifier as a string

main

public static void main(java.lang.String[] argv)

Main method for testing this class.

Parameters:

argv - the options