weka.classifiers.functions

Class XNV

java.lang.Object

weka.classifiers.AbstractClassifier

weka.classifiers.RandomizableClassifier

weka.classifiers.functions.XNV

All Implemented Interfaces:

java.io.Serializable, java.lang.Cloneable, Classifier, BatchPredictor, CapabilitiesHandler, CapabilitiesIgnorer, CommandlineRunnable, OptionHandler, Randomizable, RevisionHandler, TechnicalInformationHandler

public class XNV
extends RandomizableClassifier
implements TechnicalInformationHandler

Implements the XNV method for semi-supervised learning using a kernel function (default: RBFKernel). Standardizes all attributes, including the target, by default. Applies (unsupervised) NominalToBinary and ReplaceMissingValues before anything else is done.

For more information on the algorithm, see

Brian McWilliams, David Balduzzi, Joachim M. Buhmann: Correlated random features for fast semi-supervised learning. In: Proc 27th Annual Conference on Neural Information Processing Systems, 440--448, 2013.

BibTeX:

 @inproceedings{McWilliams2013,
    author = {Brian McWilliams and David Balduzzi and Joachim M. Buhmann},
    booktitle = {Proc 27th Annual Conference on Neural Information Processing Systems},
    pages = {440--448},
    title = {Correlated random features for fast semi-supervised learning},
    year = {2013},
    URL = {http://papers.nips.cc/paper/5000-correlated-random-features-for-fast-semi-supervised-learning.pdf}
 }
 

Valid options are:

 -S <num>
  Random number seed.
  (default 1)

 -G
  The regularization parameter gamma.

 -M
  The sample size for the Nystroem method.

 -K <kernel specification>
  The kernel function to use.

 -S
  If true, standardization will not be performed.

 -output-debug-info
  If set, classifier is run in debug mode and
  may output additional info to the console

 -do-not-check-capabilities
  If set, classifier capabilities are not checked before classifier is built
  (use with caution).

 -num-decimal-places
  The number of decimal places for the output of numbers in the model (default 2).

 -batch-size
  The desired batch size for batch prediction  (default 100).

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected no.uib.cipr.matrix.Matrix	m_B1 The CCA projection
protected boolean	m_doNotStandardize Whether to apply standardization or not.
protected double	m_Gamma The regularization parameter.
protected Kernel	m_Kernel The kernel function to use.
protected int	m_M The sample size for each Nystroem filter
protected ReplaceMissingValues	m_Missing The filter used to get rid of missing values.
protected Nystroem	m_N1 The two Nystroem filters to be used
protected Nystroem	m_N2
protected NominalToBinary	m_NominalToBinary The filter used to make attributes numeric.
protected int	m_numLabeled Stores the number of labeled instances found in the training set.
protected Standardize	m_Standardize The filter used for standardizing the data
protected no.uib.cipr.matrix.Matrix	m_wCCA The coefficients from CCA regression
protected double	m_x0
protected double	m_x1 Coefficients used compensate for standardization of the target

Fields inherited from class weka.classifiers.RandomizableClassifier

m_Seed

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

Constructor and Description
XNV()

Method Summary

Modifier and Type	Method and Description
void	buildClassifier(Instances data) Builds the XNV regressor.
static no.uib.cipr.matrix.EVD	CCA(no.uib.cipr.matrix.Matrix X1, no.uib.cipr.matrix.Matrix X2) Performs Canonical Correlation Analysis (CCA).
double[]	distributionForInstance(Instance inst) Returns prediction for an instance.
double[][]	distributionsForInstances(Instances insts) Returns predictions for a whole set of instances.
Capabilities	getCapabilities() Returns default capabilities of the classifier.
boolean	getDoNotStandardize()
double	getGamma()
Kernel	getKernel()
int	getM()
static no.uib.cipr.matrix.DenseMatrix	getMatrix(Instances data, boolean center, boolean transpose) Turns the given set of instances into a data matrix.
java.lang.String	getRevision() Returns the revision string.
TechnicalInformation	getTechnicalInformation() Returns a reference to the algorithm implemented by this class.
java.lang.String	globalInfo() Provides information regarding this class.
boolean	implementsMoreEfficientBatchPrediction() This class implements efficient batch prediction.
static no.uib.cipr.matrix.Matrix	inverse(no.uib.cipr.matrix.Matrix M) Returns the inverse of the given matrix.
static void	main(java.lang.String[] argv) Generates an XNV predictor.
void	setDoNotStandardize(boolean v)
void	setGamma(double v)
void	setKernel(Kernel kernel)
void	setM(int v)
java.lang.String	toString() Outputs the coefficients of the classifier.

Methods inherited from class weka.classifiers.RandomizableClassifier

getOptions, getSeed, listOptions, seedTipText, setOptions, setSeed

Methods inherited from class weka.classifiers.AbstractClassifier

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

Methods inherited from class java.lang.Object

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

Field Detail

m_N1

protected Nystroem m_N1

The two Nystroem filters to be used

m_N2

protected Nystroem m_N2

m_wCCA

protected no.uib.cipr.matrix.Matrix m_wCCA

The coefficients from CCA regression

m_B1

protected no.uib.cipr.matrix.Matrix m_B1

The CCA projection

m_Standardize

protected Standardize m_Standardize

The filter used for standardizing the data

m_NominalToBinary

protected NominalToBinary m_NominalToBinary

The filter used to make attributes numeric.

m_Missing

protected ReplaceMissingValues m_Missing

The filter used to get rid of missing values.

m_x1

protected double m_x1

Coefficients used compensate for standardization of the target

m_x0

protected double m_x0

m_M

protected int m_M

The sample size for each Nystroem filter

m_Kernel

protected Kernel m_Kernel

The kernel function to use.

m_Gamma

protected double m_Gamma

The regularization parameter.

m_numLabeled

protected int m_numLabeled

Stores the number of labeled instances found in the training set.

m_doNotStandardize

protected boolean m_doNotStandardize

Whether to apply standardization or not.

Constructor Detail

XNV

public XNV()

Method Detail

globalInfo

public java.lang.String globalInfo()

Provides information regarding this class.

Returns:

string describing the method that this class implements

getTechnicalInformation

public TechnicalInformation getTechnicalInformation()

Returns a reference to the algorithm implemented by this class.

Specified by:

getTechnicalInformation in interface TechnicalInformationHandler

Returns:

a reference to the algorithm implemented by this class

getGamma

@OptionMetadata(displayName="Regularization parameter gamma",
                description="The regularization parameter gamma.",
                displayOrder=1,
                commandLineParamName="G",
                commandLineParamSynopsis="-G")
public double getGamma()

setGamma

public void setGamma(double v)

getM

@OptionMetadata(displayName="Sample size for Nystroem method",
                description="The sample size for the Nystroem method.",
                displayOrder=2,
                commandLineParamName="M",
                commandLineParamSynopsis="-M")
public int getM()

setM

public void setM(int v)

setKernel

@OptionMetadata(displayName="Kernel function",
                description="The kernel function to use.",
                displayOrder=3,
                commandLineParamName="K",
                commandLineParamSynopsis="-K <kernel specification>")
public void setKernel(Kernel kernel)

getKernel

public Kernel getKernel()

getDoNotStandardize

@OptionMetadata(displayName="Do not apply standardization",
                description="If true, standardization will not be performed.",
                displayOrder=4,
                commandLineParamName="S",
                commandLineParamSynopsis="-S")
public boolean getDoNotStandardize()

setDoNotStandardize

public void setDoNotStandardize(boolean v)

getMatrix

public static no.uib.cipr.matrix.DenseMatrix getMatrix(Instances data,
                                                       boolean center,
                                                       boolean transpose)

Turns the given set of instances into a data matrix.

Parameters:

data - set of instances

center - whether to center the matrix

transpose - whether to transpose the matrix

Returns:

the matrix

inverse

public static no.uib.cipr.matrix.Matrix inverse(no.uib.cipr.matrix.Matrix M)
                                         throws java.lang.Exception

Returns the inverse of the given matrix.

Parameters:

M - the matrix to invert

Returns:

the inverse

Throws:

java.lang.Exception

CCA

public static no.uib.cipr.matrix.EVD CCA(no.uib.cipr.matrix.Matrix X1,
                                         no.uib.cipr.matrix.Matrix X2)
                                  throws java.lang.Exception

Performs Canonical Correlation Analysis (CCA).

Parameters:

X1 - the first data matrix

X2 - the second data matrix

Returns:

the eigenvalue decomposition giving the basis in X1

Throws:

java.lang.Exception

buildClassifier

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

Builds the XNV regressor.

Specified by:

buildClassifier in interface Classifier

Parameters:

data - set of instances serving as training data

Throws:

java.lang.Exception

distributionForInstance

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

Returns prediction for an instance.

Specified by:

distributionForInstance in interface Classifier

Overrides:

distributionForInstance in class AbstractClassifier

Parameters:

inst -

Returns:

a one-element array with the prediction

Throws:

java.lang.Exception - if distribution could not be computed successfully

implementsMoreEfficientBatchPrediction

public boolean implementsMoreEfficientBatchPrediction()

This class implements efficient batch prediction.

Specified by:

implementsMoreEfficientBatchPrediction in interface BatchPredictor

Overrides:

implementsMoreEfficientBatchPrediction in class AbstractClassifier

Returns:

true

distributionsForInstances

public double[][] distributionsForInstances(Instances insts)
                                     throws java.lang.Exception

Returns predictions for a whole set of instances.

Specified by:

distributionsForInstances in interface BatchPredictor

Overrides:

distributionsForInstances in class AbstractClassifier

Parameters:

insts - the instances to make predictions for

Returns:

the 2D array with results

Throws:

java.lang.Exception - if a problem occurs.

toString

public java.lang.String toString()

Outputs the coefficients of the classifier.

Overrides:

toString in class java.lang.Object

Returns:

a textual description of the classifier

getCapabilities

public Capabilities getCapabilities()

Returns default capabilities of the classifier.

Specified by:

getCapabilities in interface Classifier

Specified by:

getCapabilities in interface CapabilitiesHandler

Overrides:

getCapabilities in class AbstractClassifier

Returns:

the capabilities of this classifier

See Also:

Capabilities

getRevision

public java.lang.String getRevision()

Returns the revision string.

Specified by:

getRevision in interface RevisionHandler

Overrides:

getRevision in class AbstractClassifier

Returns:

the revision

main

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

Generates an XNV predictor.

Parameters:

argv - the options

Throws:

java.lang.Exception