weka.datagenerators.classifiers.classification

Class RandomRBF

java.lang.Object

weka.datagenerators.DataGenerator

weka.datagenerators.ClassificationGenerator

weka.datagenerators.classifiers.classification.RandomRBF

All Implemented Interfaces:

Serializable, OptionHandler, Randomizable, RevisionHandler

public class RandomRBF
extends ClassificationGenerator

RandomRBF data is generated by first creating a random set of centers for each class. Each center is randomly assigned a weight, a central point per attribute, and a standard deviation. To generate new instances, a center is chosen at random taking the weights of each center into consideration. Attribute values are randomly generated and offset from the center, where the overall vector has been scaled so that its length equals a value sampled randomly from the Gaussian distribution of the center. The particular center chosen determines the class of the instance.
RandomRBF data contains only numeric attributes as it is non-trivial to include nominal values.

Valid options are:

 -h
  Prints this help.

 -o <file>
  The name of the output file, otherwise the generated data is
  printed to stdout.

 -r <name>
  The name of the relation.

 -d
  Whether to print debug informations.

 -S
  The seed for random function (default 1)

 -n <num>
  The number of examples to generate (default 100)

 -a <num>
  The number of attributes (default 10).

 -c <num>
  The number of classes (default 2)

 -C <num>
  The number of centroids to use. (default 50)

Version:

$Revision: 1.4 $

Author:

Richard Kirkby (rkirkby at cs dot waikato dot ac dot nz), FracPete (fracpete at waikato dot ac dot nz)

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected int[]	m_centroidClasses the classes of the centroids
protected double[][]	m_centroids the centroids
protected double[]	m_centroidStdDevs the stddevs of the centroids
protected double[]	m_centroidWeights the weights of the centroids
protected int	m_NumAttributes Number of attribute the dataset should have
protected int	m_NumCentroids the number of centroids to use for generation
protected int	m_NumClasses Number of Classes the dataset should have

Fields inherited from class weka.datagenerators.ClassificationGenerator

m_NumExamples

Fields inherited from class weka.datagenerators.DataGenerator

m_CreatingRelationName, m_DatasetFormat, m_Debug, m_DefaultOutput, m_NumExamplesAct, m_OptionBlacklist, m_Output, m_Random, m_RelationName, m_Seed

Constructor Summary

Constructors

Constructor and Description
RandomRBF() initializes the generator with default values

Method Summary

Modifier and Type	Method and Description
protected int	chooseRandomIndexBasedOnProportions(double[] proportionArray, Random random) returns a random index based on the given proportions
protected int	defaultNumAttributes() returns the default number of attributes
protected int	defaultNumCentroids() returns the default number of centroids
protected int	defaultNumClasses() returns the default number of classes
Instances	defineDataFormat() Initializes the format for the dataset produced.
Instance	generateExample() Generates one example of the dataset.
Instances	generateExamples() Generates all examples of the dataset.
String	generateFinished() Generates a comment string that documentats the data generator.
String	generateStart() Generates a comment string that documentates the data generator.
int	getNumAttributes() Gets the number of attributes that should be produced.
int	getNumCentroids() Gets the number of centroids.
int	getNumClasses() Gets the number of classes the dataset should have.
String[]	getOptions() Gets the current settings of the datagenerator.
String	getRevision() Returns the revision string.
boolean	getSingleModeFlag() Return if single mode is set for the given data generator mode depends on option setting and or generator type.
String	globalInfo() Returns a string describing this data generator.
Enumeration	listOptions() Returns an enumeration describing the available options.
static void	main(String[] args) Main method for executing this class.
String	numAttributesTipText() Returns the tip text for this property
String	numCentroidsTipText() Returns the tip text for this property
String	numClassesTipText() Returns the tip text for this property
void	setNumAttributes(int numAttributes) Sets the number of attributes the dataset should have.
void	setNumCentroids(int value) Sets the number of centroids to use.
void	setNumClasses(int numClasses) Sets the number of classes the dataset should have.
void	setOptions(String[] options) Parses a list of options for this object.

Methods inherited from class weka.datagenerators.ClassificationGenerator

defaultNumExamples, getNumExamples, numExamplesTipText, setNumExamples

Methods inherited from class weka.datagenerators.DataGenerator

addToBlacklist, clearBlacklist, debugTipText, defaultNumExamplesAct, defaultOutput, defaultRelationName, defaultSeed, enumToVector, formatTipText, getDatasetFormat, getDebug, getNumExamplesAct, getOutput, getRandom, getRelationName, getRelationNameToUse, getSeed, isOnBlacklist, makeData, makeOptionString, numExamplesActTipText, outputTipText, randomTipText, relationNameTipText, removeBlacklist, runDataGenerator, seedTipText, setDatasetFormat, setDebug, setNumExamplesAct, setOutput, setRandom, setRelationName, setSeed, toStringFormat

Methods inherited from class java.lang.Object

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

Field Detail

m_NumAttributes

protected int m_NumAttributes

Number of attribute the dataset should have

m_NumClasses

protected int m_NumClasses

Number of Classes the dataset should have

m_NumCentroids

protected int m_NumCentroids

the number of centroids to use for generation

m_centroids

protected double[][] m_centroids

the centroids

m_centroidClasses

protected int[] m_centroidClasses

the classes of the centroids

m_centroidWeights

protected double[] m_centroidWeights

the weights of the centroids

m_centroidStdDevs

protected double[] m_centroidStdDevs

the stddevs of the centroids

Constructor Detail

RandomRBF

public RandomRBF()

initializes the generator with default values

Method Detail

globalInfo

public String globalInfo()

Returns a string describing this data generator.

Returns:

a description of the data generator suitable for displaying in the explorer/experimenter gui

listOptions

public Enumeration listOptions()

Returns an enumeration describing the available options.

Specified by:

listOptions in interface OptionHandler

Overrides:

listOptions in class ClassificationGenerator

Returns:

an enumeration of all the available options

setOptions

public void setOptions(String[] options)
                throws Exception

Parses a list of options for this object.

Valid options are:

 -h
  Prints this help.

 -o <file>
  The name of the output file, otherwise the generated data is
  printed to stdout.

 -r <name>
  The name of the relation.

 -d
  Whether to print debug informations.

 -S
  The seed for random function (default 1)

 -n <num>
  The number of examples to generate (default 100)

 -a <num>
  The number of attributes (default 10).

 -c <num>
  The number of classes (default 2)

 -C <num>
  The number of centroids to use. (default 50)

Specified by:

setOptions in interface OptionHandler

Overrides:

setOptions in class ClassificationGenerator

Parameters:

options - the list of options as an array of strings

Throws:

Exception - if an option is not supported

getOptions

public String[] getOptions()

Gets the current settings of the datagenerator.

Specified by:

getOptions in interface OptionHandler

Overrides:

getOptions in class ClassificationGenerator

Returns:

an array of strings suitable for passing to setOptions

See Also:

DataGenerator.removeBlacklist(String[])

defaultNumAttributes

protected int defaultNumAttributes()

returns the default number of attributes

Returns:

the default number of attributes

setNumAttributes

public void setNumAttributes(int numAttributes)

Sets the number of attributes the dataset should have.

Parameters:

numAttributes - the new number of attributes

getNumAttributes

public int getNumAttributes()

Gets the number of attributes that should be produced.

Returns:

the number of attributes that should be produced

numAttributesTipText

public String numAttributesTipText()

Returns the tip text for this property

Returns:

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

defaultNumClasses

protected int defaultNumClasses()

returns the default number of classes

Returns:

the default number of classes

setNumClasses

public void setNumClasses(int numClasses)

Sets the number of classes the dataset should have.

Parameters:

numClasses - the new number of classes

getNumClasses

public int getNumClasses()

Gets the number of classes the dataset should have.

Returns:

the number of classes the dataset should have

numClassesTipText

public String numClassesTipText()

Returns the tip text for this property

Returns:

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

defaultNumCentroids

protected int defaultNumCentroids()

returns the default number of centroids

Returns:

the default number of centroids

getNumCentroids

public int getNumCentroids()

Gets the number of centroids.

Returns:

the number of centroids.

setNumCentroids

public void setNumCentroids(int value)

Sets the number of centroids to use.

Parameters:

value - the number of centroids to use.

numCentroidsTipText

public String numCentroidsTipText()

Returns the tip text for this property

Returns:

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

getSingleModeFlag

public boolean getSingleModeFlag()
                          throws Exception

Return if single mode is set for the given data generator mode depends on option setting and or generator type.

Specified by:

getSingleModeFlag in class DataGenerator

Returns:

single mode flag

Throws:

Exception - if mode is not set yet

chooseRandomIndexBasedOnProportions

protected int chooseRandomIndexBasedOnProportions(double[] proportionArray,
                                                  Random random)

returns a random index based on the given proportions

Parameters:

proportionArray - the proportions

random - the random number generator to use

Returns:

the random index

defineDataFormat

public Instances defineDataFormat()
                           throws Exception

Initializes the format for the dataset produced. Must be called before the generateExample or generateExamples methods are used. Re-initializes the random number generator with the given seed.

Overrides:

defineDataFormat in class DataGenerator

Returns:

the format for the dataset

Throws:

Exception - if the generating of the format failed

See Also:

DataGenerator.getSeed()

generateExample

public Instance generateExample()
                         throws Exception

Generates one example of the dataset.

Specified by:

generateExample in class DataGenerator

Returns:

the generated example

Throws:

Exception - if the format of the dataset is not yet defined

Exception - if the generator only works with generateExamples which means in non single mode

generateExamples

public Instances generateExamples()
                           throws Exception

Generates all examples of the dataset. Re-initializes the random number generator with the given seed, before generating instances.

Specified by:

generateExamples in class DataGenerator

Returns:

the generated dataset

Throws:

Exception - if the format of the dataset is not yet defined

Exception - if the generator only works with generateExample, which means in single mode

See Also:

DataGenerator.getSeed()

generateStart

public String generateStart()

Generates a comment string that documentates the data generator. By default this string is added at the beginning of the produced output as ARFF file type, next after the options.

Specified by:

generateStart in class DataGenerator

Returns:

string contains info about the generated rules

generateFinished

public String generateFinished()
                        throws Exception

Generates a comment string that documentats the data generator. By default this string is added at the end of theproduces output as ARFF file type.

Specified by:

generateFinished in class DataGenerator

Returns:

string contains info about the generated rules

Throws:

Exception - if the generating of the documentaion fails

getRevision

public String getRevision()

Returns the revision string.

Returns:

the revision

main

public static void main(String[] args)

Main method for executing this class.

Parameters:

args - should contain arguments for the data producer: