keel.Algorithms.Instance_Generation.Basic

Class PrototypeGenerator

java.lang.Object

keel.Algorithms.Instance_Generation.Basic.PrototypeGenerator

Direct Known Subclasses:

AMPSOGenerator, AVG, AVQGenerator, BTS3Generator, ChenGenerator, CNN, DEGenerator, DEGLGenerator, ENPCGenerator, GENNGenerator, HYBGenerator, ICPLGenerator, IPLDEGenerator, JADEGenerator, LVQGenerator, LVQPRU, MixtGaussGenerator, MSEGenerator, OBDEGenerator, PNNGenerator, POCGenerator, PSCSAGenerator, PSOGenerator, RandomSelector, RSPGenerator, SADEGenerator, SFLSDEGenerator, SGPGenerator, SVMSEL

public class PrototypeGenerator
extends java.lang.Object

Implements a generic Prototype Generator

Author:

diegoj

Field Summary

Fields

Modifier and Type	Field and Description
protected java.lang.String	algorithmName Name of the reduction tecnique
protected PrototypeSet	generatedDataSet Condensed data
static long	SEED Default seed value to the Random Number Generator
long[]	seedDefaultValueList Default seed list to the Random Number Generator
protected PrototypeSet	trainingDataSet Original data set to be condensed

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

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.
void	desordenar_vector_sin(int[] vector) Shuffles the values of the vector given as parameter, maintaining the last element value.
void	desordenar_vector(int[] vector) Shuffles the values of the vector given as parameter.
PrototypeSet	execute() Execute the reduction of the data set.
PrototypeSet	generateReducedDataSet() Execute the reduction of the data set
java.lang.String	getResultingAccuracy(java.lang.String name, int accuracy1NN, PrototypeSet test) Internal function that gets only the accuracy of the condensation.
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.
java.lang.String	getResults(java.lang.String name, java.lang.String algoName, int accuracy1NN, int training_size, PrototypeSet test) Internal function that gets the parameters of accuracy of the condensation.
java.lang.String	getResultsOfAccuracy(java.lang.String name, int accuracy1NN, PrototypeSet test) Internal function that gets the parameters of 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) Sets each position of the vector given as parameter with the i-th, skipping the position given as argument.
void	inic_vector(int[] vector) Sets each position of the vector given as parameter with the i-th number.
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
PrototypeSet	reduceSet() Makes the trivial reduction.
void	saveResultsOfAccuracyIn(java.lang.String name, int accuracy1NN, PrototypeSet test, java.lang.String fileName) Internal function that shows in the screen the parameters of accuracy of the condensation
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	setSeed(long _seed) Set the seed of the random generator.
void	showResultsOfAccuracy(int accuracyKNN, int accuracy1NN, int k, PrototypeSet test) Internal function that shows in the screen the parameters of accuracy of the condensation
void	showResultsOfAccuracy(java.lang.String name, int accuracy1NN, PrototypeSet test) Internal function that shows in the screen the parameters of accuracy of the condensation

Methods inherited from class java.lang.Object

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

Field Detail

trainingDataSet

protected PrototypeSet trainingDataSet

Original data set to be condensed

generatedDataSet

protected 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])

Method Detail

setSeed

public static void setSeed(long _seed)

Set the seed of the random generator.

Parameters:

_seed - Seed value.

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.

reduceSet

public PrototypeSet reduceSet()

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

Returns:

A copy of the training data set

execute

public final PrototypeSet execute()

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

Returns:

Reduced data set.

generateReducedDataSet

public PrototypeSet generateReducedDataSet()

Execute the reduction of the data set

Returns:

Reduced data set.

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 neighbour

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 classified and number of the not well classified

showResultsOfAccuracy

public void showResultsOfAccuracy(int accuracyKNN,
                                  int accuracy1NN,
                                  int k,
                                  PrototypeSet test)

Internal function that shows in the screen the parameters of accuracy of the condensation

Parameters:

accuracyKNN - Number of well-classificated prototypes with KNN.

accuracy1NN - Number of well-classificated prototypes with 1NN.

k - Number of neighbors in the KNN.

test - Test prototype set.

getResultsOfAccuracy

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

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

Parameters:

name - Name of the data set.

accuracy1NN - Number of well-classificated prototypes with KNN.

test - Test prototype set.

Returns:

Return the results of the accuracy.

getResults

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

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

Parameters:

name - Name of the data set.

algoName - Name of the algorithm.

accuracy1NN - Number of well-classificated prototypes with KNN.

training_size - Training dataset size (number of examples).

test - Test prototype set.

Returns:

Return the results of the accuracy.

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.

getResultingAccuracy

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

Internal function that gets only the accuracy of the condensation.

Parameters:

name - Name of the data set.

accuracy1NN - Number of well-classificated prototypes with KNN.

test - Test prototype set.

Returns:

Return the results of the accuracy.

showResultsOfAccuracy

public void showResultsOfAccuracy(java.lang.String name,
                                  int accuracy1NN,
                                  PrototypeSet test)

Internal function that shows in the screen the parameters of accuracy of the condensation

Parameters:

name - Name of the data set.

accuracy1NN - Number of well-classificated prototypes with KNN.

test - Test prototype set.

saveResultsOfAccuracyIn

public void saveResultsOfAccuracyIn(java.lang.String name,
                                    int accuracy1NN,
                                    PrototypeSet test,
                                    java.lang.String fileName)

Internal function that shows in the screen the parameters of accuracy of the condensation

Parameters:

name - Name of the data set.

accuracy1NN - Number of well-classificated prototypes with KNN.

test - Test prototype set.

fileName - Output file.

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)

Sets each position of the vector given as parameter with the i-th number.

Parameters:

vector - vector to complete.

inic_vector_sin

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

Sets each position of the vector given as parameter with the i-th, skipping the position given as argument.

Parameters:

vector - vector to complete.

without - position to skip.

desordenar_vector_sin

public void desordenar_vector_sin(int[] vector)

Shuffles the values of the vector given as parameter, maintaining the last element value.

Parameters:

vector - vector to shuffle.

desordenar_vector

public void desordenar_vector(int[] vector)

Shuffles the values of the vector given as parameter.

Parameters:

vector - vector to shuffle.

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.