keel.Algorithms.Semi_Supervised_Learning.Basic

Class PrototypeGenerator

java.lang.Object

keel.Algorithms.Semi_Supervised_Learning.Basic.PrototypeGenerator

Direct Known Subclasses:

ADE_CoForestGenerator, APSSCGenerator, C45SSLGenerator, CLCCGenerator, CoBCGenerator, CoForestGenerator, CoTrainingGenerator, DE_TriTrainingGenerator, DemocraticGenerator, NBSSLGenerator, NNSSLGenerator, RASCOGenerator, Rel_RASCOGenerator, SelfTrainingGenerator, SETREDGenerator, SMOSSLGenerator, SNNRCEGenerator, TriTrainingGenerator

public class PrototypeGenerator
extends java.lang.Object

Implements a generic Semi-supervised method

Author:

Isaac Triguero

Field Summary

Fields

Modifier and Type	Field and Description
protected java.lang.String	algorithmName Name of the reduction tecnique
protected Pair<PrototypeSet,PrototypeSet>	generatedDataSet Condensed data
static InstanceSet	Instancetest Original test data set to be condensed
static InstanceSet	Instancetrain Original training data set to be condensed
static long	SEED Default seed value to the Random Number Generator
long[]	seedDefaultValueList Default seed list to the Random Number Generator
protected PrototypeSet	testDataSet Test prototype set.
protected PrototypeSet	trainingDataSet Training prototype set.
protected PrototypeSet	transductiveDataSet Transductive prototype set.

Constructor Summary

Constructors

Constructor and Description
PrototypeGenerator(PrototypeSet _trainingDataSet) Construct the PrototypeGenerator
PrototypeGenerator(PrototypeSet _trainingDataSet, int seedIndex) Construct the PrototypeGenerator
PrototypeGenerator(PrototypeSet _trainingDataSet, Parameters parameters) Construct the PrototypeGenerator
PrototypeGenerator(PrototypeSet _trainingDataSet, PrototypeSet _testDataSet, Parameters parameters) Construct the PrototypeGenerator for Supervised Learning (pos-processing)
PrototypeGenerator(PrototypeSet _trainingDataSet, PrototypeSet _unlabeledDataSet, PrototypeSet _testDataSet, Parameters parameters) Construct the PrototypeGenerator

Method Summary

Modifier and Type	Method and Description
protected static int	absoluteAccuracy(PrototypeSet condensed, PrototypeSet test) Calculate the absolute accuracy between two sets
protected static Pair<java.lang.Integer,java.lang.Integer>	absoluteAccuracyAndError(PrototypeSet condensed, PrototypeSet test) Calculate the absolute accuracy between two sets
protected static int	absoluteAccuracyKNN(PrototypeSet condensed, PrototypeSet test, int k) Calculate the absolute accuracy between two sets
protected static double	accuracy(PrototypeSet condensed, PrototypeSet test) Calculate the percetage of well classificated prototypes of one set using 1NN in a reduced set.
double	accuracy2(PrototypeSet condensed, PrototypeSet test) Calculate the percetage of well classificated prototypes of one set using 1NN in a reduced set.
Pair<PrototypeSet,PrototypeSet>	applyAlgorithm() Makes the trivial reduction.
void	desordenar_vector_sin(int[] vector) Cuando quitas uno, con el inic vector, el desordenar no puede coger el ultimo..
void	desordenar_vector(int[] vector) Shuffles the vector given.
Pair<PrototypeSet,PrototypeSet>	execute() Execute the reduction of the data set.
Pair<PrototypeSet,PrototypeSet>	generateReducedDataSet() Execute the reduction of the data set
java.lang.String	getResultingAccuracy(java.lang.String name, java.lang.String algoName, int accuracy1NN, PrototypeSet test) Internal function that gets only the accuracy of the condensation.
static java.lang.String	getResultsOfAccuracy(java.lang.String name, java.lang.String algorithmUsed, PrototypeSet reduced, PrototypeSet test) Internal function that gets the parameters of accuracy of the condensation.
static long	getSeed() Get the seed of the random generator.
protected int	getSetSizeFromPercentage(double percentage) Number of prototypes corresponding of the desired percentage of the reduced set.
protected static int	getSetSizeFromPercentage(PrototypeSet set, double percentage) Number of prototypes corresponding of the desired percentage of the reduced set.
long	getTime() Tell the time elapsed.
void	inic_vector_sin(int[] vector, int without) Initiates the vector given without the index given.
void	inic_vector(int[] vector) Initiates the vector given.
static void	main(java.lang.String[] args) General main for all the prototoype generators Arguments: 0: Filename with the training data set to be condensed. 1: Filename wich will contain the test data set 3: k Number of neighbors used in the KNN function
static void	saveResultsOfAccuracyIn(java.lang.String name, java.lang.String algorithmUsed, PrototypeSet reduced, PrototypeSet test, java.lang.String fileName, boolean append) Internal function that saves the parameters of accuracy of the condensation.
PrototypeSet	selecRandomSet(int numberOfPrototypesSelected, boolean usePriorProb) Extract a random prototype set of the original traning data set.
static void	setInstanceTest(InstanceSet uno) Sets the test instances set given.
static void	setInstanceTrain(InstanceSet uno) Sets the training instances set given.
static void	setSeed(long _seed) Set the seed of the random generator.

Methods inherited from class java.lang.Object

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

Field Detail

Instancetrain

public static InstanceSet Instancetrain

Original training data set to be condensed

Instancetest

public static InstanceSet Instancetest

Original test data set to be condensed

trainingDataSet

protected PrototypeSet trainingDataSet

Training prototype set.

transductiveDataSet

protected PrototypeSet transductiveDataSet

Transductive prototype set.

testDataSet

protected PrototypeSet testDataSet

Test prototype set.

generatedDataSet

protected Pair<PrototypeSet,PrototypeSet> generatedDataSet

Condensed data

algorithmName

protected java.lang.String algorithmName

Name of the reduction tecnique

SEED

public static long SEED

Default seed value to the Random Number Generator

seedDefaultValueList

public final long[] seedDefaultValueList

Default seed list to the Random Number Generator

Constructor Detail

PrototypeGenerator

public PrototypeGenerator(PrototypeSet _trainingDataSet)

Construct the PrototypeGenerator

Parameters:

_trainingDataSet - Original data to be condensed.

PrototypeGenerator

public PrototypeGenerator(PrototypeSet _trainingDataSet,
                          int seedIndex)

Construct the PrototypeGenerator

Parameters:

_trainingDataSet - Original data to be condensed.

seedIndex - Index of the seedDefaultValueList

PrototypeGenerator

public PrototypeGenerator(PrototypeSet _trainingDataSet,
                          Parameters parameters)

Construct the PrototypeGenerator

Parameters:

_trainingDataSet - Original data to be condensed.

parameters - Parameters of the algorithm (the random seedDefaultValueList in [0])

PrototypeGenerator

public PrototypeGenerator(PrototypeSet _trainingDataSet,
                          PrototypeSet _testDataSet,
                          Parameters parameters)

Construct the PrototypeGenerator for Supervised Learning (pos-processing)

Parameters:

_trainingDataSet - Original data to be condensed.

_testDataSet - Original test prototype set.

parameters - Parameters of the algorithm (the random seedDefaultValueList in [0])

PrototypeGenerator

public PrototypeGenerator(PrototypeSet _trainingDataSet,
                          PrototypeSet _unlabeledDataSet,
                          PrototypeSet _testDataSet,
                          Parameters parameters)

Construct the PrototypeGenerator

Parameters:

_trainingDataSet - Original data to be condensed.

_unlabeledDataSet - Original unlabeled prototype set for SSL.

_testDataSet - Original test prototype set.

parameters - Parameters of the algorithm (the random seedDefaultValueList in [0])

Method Detail

setSeed

public static void setSeed(long _seed)

Set the seed of the random generator.

Parameters:

_seed - Seed value.

setInstanceTrain

public static void setInstanceTrain(InstanceSet uno)

Sets the training instances set given.

Parameters:

uno - the training instances set to set.

setInstanceTest

public static void setInstanceTest(InstanceSet uno)

Sets the test instances set given.

Parameters:

uno - the test instances set to set.

getSeed

public static long getSeed()

Get the seed of the random generator.

Returns:

Seed value of the random generator.

getTime

public long getTime()

Tell the time elapsed.

Returns:

Time elapsed.

getSetSizeFromPercentage

protected int getSetSizeFromPercentage(double percentage)

Number of prototypes corresponding of the desired percentage of the reduced set.

Parameters:

percentage - Percentage of size that will have reduced set.

Returns:

Number of prototypes corresponding of the desired percentage of the reduced set.

getSetSizeFromPercentage

protected static int getSetSizeFromPercentage(PrototypeSet set,
                                              double percentage)

Number of prototypes corresponding of the desired percentage of the reduced set.

Parameters:

set - Original prototype set.

percentage - Percentage of size that will have reduced set.

Returns:

Number of prototypes corresponding of the desired percentage of the reduced set.

selecRandomSet

public PrototypeSet selecRandomSet(int numberOfPrototypesSelected,
                                   boolean usePriorProb)

Extract a random prototype set of the original traning data set.

Parameters:

numberOfPrototypesSelected - Number of prototypes extracted.

usePriorProb - Use the a priori probabilites of the set.

Returns:

Set which have the specified number of prototypes randomly selected.

applyAlgorithm

public Pair<PrototypeSet,PrototypeSet> applyAlgorithm()
                                               throws java.lang.Exception

Makes the trivial reduction. That is none reduction of the set

Returns:

A copy of the training data set

Throws:

java.lang.Exception - if the algorithm can not be applied.

execute

public final Pair<PrototypeSet,PrototypeSet> execute()
                                              throws java.lang.Exception

Execute the reduction of the data set. Note that this function, au contraire that reduceSet, uses timers.

Returns:

Reduced data set.

Throws:

java.lang.Exception - if the algorithm can not be executed.

generateReducedDataSet

public Pair<PrototypeSet,PrototypeSet> generateReducedDataSet()
                                                       throws java.lang.Exception

Execute the reduction of the data set

Returns:

Reduced data set.

Throws:

java.lang.Exception - if the reduction can not be done.

absoluteAccuracy

protected static int absoluteAccuracy(PrototypeSet condensed,
                                      PrototypeSet test)

Calculate the absolute accuracy between two sets

Parameters:

condensed - Reduced data set

test - Test data set

Returns:

Number of prototypes of test set that has been well classificated with 1NN on condensed.

absoluteAccuracyKNN

protected static int absoluteAccuracyKNN(PrototypeSet condensed,
                                         PrototypeSet test,
                                         int k)

Calculate the absolute accuracy between two sets

Parameters:

condensed - Reduced data set

test - Test data set

k - number of nearest neighbours considered.

Returns:

Number of prototypes of test set that has been well classificate with KNN on condensed.

accuracy

protected static double accuracy(PrototypeSet condensed,
                                 PrototypeSet test)

Calculate the percetage of well classificated prototypes of one set using 1NN in a reduced set.

Parameters:

condensed - Reduced data set

test - Test data set

Returns:

Percetage of well classificated prototypes of test in condendesd set.

accuracy2

public double accuracy2(PrototypeSet condensed,
                        PrototypeSet test)

Calculate the percetage of well classificated prototypes of one set using 1NN in a reduced set.

Parameters:

condensed - Reduced data set

test - Test data set

Returns:

Percetage of well classificated prototypes of test in condendesd set.

absoluteAccuracyAndError

protected static Pair<java.lang.Integer,java.lang.Integer> absoluteAccuracyAndError(PrototypeSet condensed,
                                                                                    PrototypeSet test)

Calculate the absolute accuracy between two sets

Parameters:

condensed - Reduced data set

test - Test data set

Returns:

Number of prototypes well classificated

getResultingAccuracy

public java.lang.String getResultingAccuracy(java.lang.String name,
                                             java.lang.String algoName,
                                             int accuracy1NN,
                                             PrototypeSet test)

Internal function that gets only the accuracy of the condensation.

Parameters:

name - Name of the data set.

algoName - Algorithm name.

accuracy1NN - Number of well-classificated prototypes with KNN.

test - Test prototype set.

Returns:

Return the results of the accuracy.

getResultsOfAccuracy

public static java.lang.String getResultsOfAccuracy(java.lang.String name,
                                                    java.lang.String algorithmUsed,
                                                    PrototypeSet reduced,
                                                    PrototypeSet test)

Internal function that gets the parameters of accuracy of the condensation.

Parameters:

name - Name of the data set.

algorithmUsed - Name of the algorithm used to reduce the set.

reduced - Reduced prototype set.

test - Test prototype set.

Returns:

Return the results of the accuracy as string.

saveResultsOfAccuracyIn

public static void saveResultsOfAccuracyIn(java.lang.String name,
                                           java.lang.String algorithmUsed,
                                           PrototypeSet reduced,
                                           PrototypeSet test,
                                           java.lang.String fileName,
                                           boolean append)

Internal function that saves the parameters of accuracy of the condensation.

Parameters:

name - Name of the data set.

algorithmUsed - Name of the algorithm used to reduce the set.

reduced - Reduced prototype set.

test - Test prototype set.

fileName - File name that will contain the accuracy.

append - If is TRUE, appends the results to the output file; if is false rewrites the file with the new data.

inic_vector

public void inic_vector(int[] vector)

Initiates the vector given.

Parameters:

vector - given vector to initiates

inic_vector_sin

public void inic_vector_sin(int[] vector,
                            int without)

Initiates the vector given without the index given.

Parameters:

vector - given vector to initiates

without - index given.

desordenar_vector_sin

public void desordenar_vector_sin(int[] vector)

Cuando quitas uno, con el inic vector, el desordenar no puede coger el ultimo.. necesito otro metodo

Parameters:

vector -

desordenar_vector

public void desordenar_vector(int[] vector)

Shuffles the vector given.

Parameters:

vector - given vector to shuffles.

main

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

General main for all the prototoype generators Arguments: 0: Filename with the training data set to be condensed. 1: Filename wich will contain the test data set 3: k Number of neighbors used in the KNN function

Parameters:

args - Arguments of the main function.