keel.Algorithms.Instance_Generation.SSMAPSO

Class Cromosoma

java.lang.Object

keel.Algorithms.Instance_Generation.SSMAPSO.Cromosoma

All Implemented Interfaces:

java.lang.Comparable

public class Cromosoma
extends java.lang.Object
implements java.lang.Comparable

File: Cromosoma.java Auxiliriary class to represent chromosomes for Instance Generation methods

Since:

JDK1.5

Version:

0.1

Author:

Written by Salvador García (University of Granada) 20/07/2004

Constructor Summary

Constructors

Constructor and Description
Cromosoma(int K, Cromosoma a, Cromosoma b, double pCross, int size) Construct a chromosome throught crossover than other two parents
Cromosoma(int K, int size, Cromosoma a) Create a copied chromosome
Cromosoma(int K, int size, double[][] dMatrix, double[][] datos, double[][] real, int[][] nominal, boolean[][] nulo, boolean distanceEu) Construct a random chromosome of specified size and evaluate it.

Method Summary

Modifier and Type	Method and Description
void	borrar() Marks a chromosome for deletion.
int	compareTo(java.lang.Object o1) Compare to Method
int	differenceAtOne(Cromosoma a) Test if two chromosome differ in only one gene
boolean	estaEvaluado() Tests if the chromosome is already evaluated
boolean	esValido() Tests if the chromosome is valid
void	evaluacionCompleta(int nClases, int K, int[] clases) Function that evaluates a cromosome completely
double	evaluacionParcial(int nClases, int K, int[] clases, int ref, double[][] dMatrix, double umbral, double[][] datos, double[][] real, int[][] nominal, boolean[][] nulo, boolean distanceEu) Partial evaluation.
int	genesActivos() Count the number of genes set to 1
double	getFitness() Get the fitness of a chromosome
double	getFitnessAc() Get the accuracy fitness of a chromosome
boolean	getGen(int indice) Get the value of a gene
int	getVecino(int indicei, int indicej) Get the neighbour with the given indeces.
void	mutation(int K, double pMut, double[][] dMatrix, double[][] datos, double[][] real, int[][] nominal, boolean[][] nulo, boolean distanceEu) Function that does the mutation
int	obtenerCercano(int[] vecinos, int J, double[][] dMatrix, int index, double[][] datos, double[][] real, int[][] nominal, boolean[][] nulo, boolean distanceEu) Obtains the nearest neighbour given a mask (chromosome)
double	optimizacionLocal(int nClases, int K, int[] clases, double[][] dMatrix, double umbral, double[][] datos, double[][] real, int[][] nominal, boolean[][] nulo, boolean distanceEu) Local optimization for the chromosome as Memetic algorithm
void	setGen(int pos, boolean valor) Set the value of a gene
java.lang.String	toString() To String Method

Methods inherited from class java.lang.Object

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

Constructor Detail

Cromosoma

public Cromosoma(int K,
                 int size,
                 double[][] dMatrix,
                 double[][] datos,
                 double[][] real,
                 int[][] nominal,
                 boolean[][] nulo,
                 boolean distanceEu)

Construct a random chromosome of specified size and evaluate it.

Parameters:

K - Number of neighbors

size - Size of the chromosome

dMatrix - data matrix.

datos - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulo - Reference to the training set (null values)

distanceEu - True= Euclidean distance; False= HVDM

Cromosoma

public Cromosoma(int K,
                 int size,
                 Cromosoma a)

Create a copied chromosome

Parameters:

K - Number of neighbors

size - Chromosome size.

a - Chromosome to copy.

Cromosoma

public Cromosoma(int K,
                 Cromosoma a,
                 Cromosoma b,
                 double pCross,
                 int size)

Construct a chromosome throught crossover than other two parents

Parameters:

K - Number of neighbors

a - first father

b - second father.

pCross - crossover probability

size - Chromosome size.

Method Detail

mutation

public void mutation(int K,
                     double pMut,
                     double[][] dMatrix,
                     double[][] datos,
                     double[][] real,
                     int[][] nominal,
                     boolean[][] nulo,
                     boolean distanceEu)

Function that does the mutation

Parameters:

K - Number of neighbors

pMut - Mutation probability.

dMatrix - data matrix.

datos - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulo - Reference to the training set (null values)

distanceEu - True= Euclidean distance; False= HVDM

obtenerCercano

public int obtenerCercano(int[] vecinos,
                          int J,
                          double[][] dMatrix,
                          int index,
                          double[][] datos,
                          double[][] real,
                          int[][] nominal,
                          boolean[][] nulo,
                          boolean distanceEu)

Obtains the nearest neighbour given a mask (chromosome)

Parameters:

vecinos - the nearest neighbour

J - chromosome size.

dMatrix - data matrix.

index - example index.

datos - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulo - Reference to the training set (null values)

distanceEu - True= Euclidean distance; False= HVDM

Returns:

minimum position of the nearest neighbour of the vector of neighbours

getGen

public boolean getGen(int indice)

Get the value of a gene

Parameters:

indice - Index of the gene

Returns:

Value of the especified gene

getVecino

public int getVecino(int indicei,
                     int indicej)

Get the neighbour with the given indeces.

Parameters:

indicei - gene index.

indicej - neighbour index

Returns:

the neighbour with the given indeces.

getFitness

public double getFitness()

Get the fitness of a chromosome

Returns:

fitness of the chromosome

getFitnessAc

public double getFitnessAc()

Get the accuracy fitness of a chromosome

Returns:

accuracy fitness of the chromosome

evaluacionCompleta

public void evaluacionCompleta(int nClases,
                               int K,
                               int[] clases)

Function that evaluates a cromosome completely

Parameters:

nClases - number of classes

K - Number of neighbors

clases - Output attribute of each instance

borrar

public void borrar()

Marks a chromosome for deletion.

esValido

public boolean esValido()

Tests if the chromosome is valid

Returns:

True if the chromosome is valid. False, if not.

setGen

public void setGen(int pos,
                   boolean valor)

Set the value of a gene

Parameters:

pos - Index of the gene

valor - Value to set

estaEvaluado

public boolean estaEvaluado()

Tests if the chromosome is already evaluated

Returns:

True if the chromosome is already evaluated. False, if not.

genesActivos

public int genesActivos()

Count the number of genes set to 1

Returns:

Number of genes set to 1 in the chromosome

optimizacionLocal

public double optimizacionLocal(int nClases,
                                int K,
                                int[] clases,
                                double[][] dMatrix,
                                double umbral,
                                double[][] datos,
                                double[][] real,
                                int[][] nominal,
                                boolean[][] nulo,
                                boolean distanceEu)

Local optimization for the chromosome as Memetic algorithm

Parameters:

nClases - number of classes

K - Number of neighbors

clases - Output attribute of each instance

dMatrix - data matrix.

umbral - Threshold. * @param vecinos the nearest neighbours

datos - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulo - Reference to the training set (null values)

distanceEu - True= Euclidean distance; False= HVDM

Returns:

number of evaluations

evaluacionParcial

public double evaluacionParcial(int nClases,
                                int K,
                                int[] clases,
                                int ref,
                                double[][] dMatrix,
                                double umbral,
                                double[][] datos,
                                double[][] real,
                                int[][] nominal,
                                boolean[][] nulo,
                                boolean distanceEu)

Partial evaluation.

Parameters:

nClases - number of classes

K - Number of neighbors

clases - Output attribute of each instance

ref - reference id.

dMatrix - data matrix.

umbral - Threshold. * @param vecinos the nearest neighbours

datos - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulo - Reference to the training set (null values)

distanceEu - True= Euclidean distance; False= HVDM

Returns:

partial evaluations.

compareTo

public int compareTo(java.lang.Object o1)

Compare to Method

Specified by:

compareTo in interface java.lang.Comparable

Parameters:

o1 - Chromosome to compare

Returns:

Relative order between the chromosomes

differenceAtOne

public int differenceAtOne(Cromosoma a)

Test if two chromosome differ in only one gene

Parameters:

a - Chromosome to compare

Returns:

Position of the difference, if only one is found. Otherwise, -1

toString

public java.lang.String toString()

To String Method

Overrides:

toString in class java.lang.Object

Returns:

String representation of the chromosome