keel.Algorithms.Decision_Trees.C45_Binarization

Class KNN

java.lang.Object

keel.Algorithms.Decision_Trees.C45_Binarization.KNN

public class KNN
extends java.lang.Object

File: KNN.java An auxiliary implementation of the KNN classifier for using in Instance Selection algorithms

Since:

JDK1.5

Version:

0.2

Author:

Written by Salvador García (University of Granada) 20/07/2004, Modified by Isaac Triguero (University of Granada) 20/06/2010

Constructor Summary

Constructors

Constructor and Description
KNN()

Method Summary

Modifier and Type	Method and Description
protected static double	distance(double[] instance1, double[] instance2) Calculates the Euclidean distance between two instances
static double	distancia(double[] ej1, double[] ej2) Calculates the Euclidean distance between two instances
static double	distancia(double[] ej1, double[] ej1Real, int[] ej1Nom, boolean[] ej1Nul, double[] ej2, double[] ej2Real, int[] ej2Nom, boolean[] ej2Nul, boolean Euc) Calculates the HVDM distance between two instances
static double	distancia2(double[] ej1, double[] ej2) Calculates the unsquared Euclidean distance between two instances
static double	distancia2(double[] ej1, double[] ej1Real, int[] ej1Nom, boolean[] ej1Nul, double[] ej2, double[] ej2Real, int[] ej2Nom, boolean[] ej2Nul, boolean Euc) Calculates the unsquared HVDM distance between two instances
static int	evaluacionKNN(int nvec, double[][] conj, double[][] real, int[][] nominal, boolean[][] nulos, int[] clases, double[] ejemplo, double[] ejReal, int[] ejNominal, boolean[] ejNulos, int nClases, boolean distance) Executes KNN
static int	evaluacionKNN(int nvec, double[][] conj, int[] clases, double[] ejemplo, int nClases) Executes KNN
static int	evaluacionKNN2(int nvec, double[][] conj, double[][] real, int[][] nominal, boolean[][] nulos, int[] clases, double[] ejemplo, double[] ejReal, int[] ejNominal, boolean[] ejNulos, int nClases, boolean distance) Executes KNN
static int	evaluacionKNN2(int nvec, double[][] conj, double[][] real, int[][] nominal, boolean[][] nulos, int[] clases, double[] ejemplo, double[] ejReal, int[] ejNominal, boolean[] ejNulos, int nClases, boolean distance, int[] vecinos) Executes KNN
static int	evaluacionKNN2(int nvec, double[][] conj, double[][] real, int[][] nominal, boolean[][] nulos, int[] clases, double[] ejemplo, double[] ejReal, int[] ejNominal, boolean[] ejNulos, int nClases, boolean distance, int[] vecinos, double[] minDistanciasSal, int[] votosSal) Executes KNN
static int	evaluacionKNN2(int nvec, double[][] conj, double[][] real, int[][] nominal, boolean[][] nulos, int[] clases, double[] ejemplo, double[] ejReal, int[] ejNominal, boolean[] ejNulos, int nClases, boolean distance, Referencia nVotos) Executes KNN
static int	evaluacionKNN2(int nvec, double[][] conj, int[] clases, double[] ejemplo, int nClases) Executes KNN
static int	evaluacionKNN2(int nvec, double[][] conj, int[] clases, double[] ejemplo, int nClases, int[] vecinos) Executes KNN
static int	evaluacionKNN2(int nvec, double[][] conj, int[] clases, double[] ejemplo, int nClases, Referencia nVotos) Executes KNN
static int[]	evaluacionKNN3(int nvec, double[][] conj, double[][] real, int[][] nominal, boolean[][] nulos, int[] clases, double[] ejemplo, double[] ejReal, int[] ejNominal, boolean[] ejNulos, int nClases, boolean distance) Executes KNN
static int[]	evaluacionKNN3(int nvec, double[][] conj, int[] clases, double[] ejemplo, int nClases) Executes KNN
static int	evaluate(double[] example, double[][] trainData, int nClasses, int[] trainOutput, int k) Evaluates a instance to predict its class.
static void	writeOutput(java.lang.String filename, int[][] realClass, int[][] prediction, Attribute[] inputs, Attribute output, java.lang.String relation) Prints output files.

Methods inherited from class java.lang.Object

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

Constructor Detail

KNN

public KNN()

Method Detail

writeOutput

public static void writeOutput(java.lang.String filename,
                               int[][] realClass,
                               int[][] prediction,
                               Attribute[] inputs,
                               Attribute output,
                               java.lang.String relation)

Prints output files.

Parameters:

filename - Name of output file

realClass - Real output of instances

prediction - Predicted output for instances

inputs - Inputs attributes

output - Output attribute

relation - String relation

distance

protected static double distance(double[] instance1,
                                 double[] instance2)

Calculates the Euclidean distance between two instances

Parameters:

instance1 - First instance

instance2 - Second instance

Returns:

The Euclidean distance

evaluate

public static int evaluate(double[] example,
                           double[][] trainData,
                           int nClasses,
                           int[] trainOutput,
                           int k)

Evaluates a instance to predict its class.

Parameters:

example - Instance evaluated

trainData - Training input data

nClasses - Number of classes

trainOutput - Training output data

k - Number of neighbors considered

Returns:

Class predicted

evaluacionKNN

public static int evaluacionKNN(int nvec,
                                double[][] conj,
                                int[] clases,
                                double[] ejemplo,
                                int nClases)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

clases - Output attribute of each instance

ejemplo - New instance to classifiy

nClases - Number of classes of the problem

Returns:

Class of the new instance

evaluacionKNN2

public static int evaluacionKNN2(int nvec,
                                 double[][] conj,
                                 int[] clases,
                                 double[] ejemplo,
                                 int nClases)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

clases - Output attribute of each instance

ejemplo - New instance to classifiy

nClases - Number of classes of the problem

Returns:

Class of the new instance

evaluacionKNN3

public static int[] evaluacionKNN3(int nvec,
                                   double[][] conj,
                                   int[] clases,
                                   double[] ejemplo,
                                   int nClases)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

clases - Output attribute of each instance

ejemplo - New instance to classifiy

nClases - Number of classes of the problem

Returns:

The neighboors' classes.

evaluacionKNN2

public static int evaluacionKNN2(int nvec,
                                 double[][] conj,
                                 int[] clases,
                                 double[] ejemplo,
                                 int nClases,
                                 Referencia nVotos)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

clases - Output attribute of each instance

ejemplo - New instance to classifiy

nClases - Number of classes of the problem

nVotos - Maximun number of votes achieved

Returns:

Class of the new instance

evaluacionKNN2

public static int evaluacionKNN2(int nvec,
                                 double[][] conj,
                                 int[] clases,
                                 double[] ejemplo,
                                 int nClases,
                                 int[] vecinos)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

clases - Output attribute of each instance

ejemplo - New instance to classifiy

nClases - Number of classes of the problem

vecinos - Neighbors of the new instance

Returns:

Class of the new instance

distancia

public static double distancia(double[] ej1,
                               double[] ej2)

Calculates the Euclidean distance between two instances

Parameters:

ej1 - First instance

ej2 - Second instance

Returns:

The Euclidean distance

distancia2

public static double distancia2(double[] ej1,
                                double[] ej2)

Calculates the unsquared Euclidean distance between two instances

Parameters:

ej1 - First instance

ej2 - Second instance

Returns:

The unsquared Euclidean distance

evaluacionKNN

public static int evaluacionKNN(int nvec,
                                double[][] conj,
                                double[][] real,
                                int[][] nominal,
                                boolean[][] nulos,
                                int[] clases,
                                double[] ejemplo,
                                double[] ejReal,
                                int[] ejNominal,
                                boolean[] ejNulos,
                                int nClases,
                                boolean distance)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulos - Reference to the training set (null values)

clases - Output attribute of each instance

ejemplo - New instance to classifiy

ejReal - New instance to classifiy (real valued)

ejNominal - New instance to classifiy (nominal valued)

ejNulos - New instance to classifiy (null values)

nClases - Number of classes of the problem

distance - True= Euclidean distance; False= HVDM

Returns:

Class of the new instance

evaluacionKNN2

public static int evaluacionKNN2(int nvec,
                                 double[][] conj,
                                 double[][] real,
                                 int[][] nominal,
                                 boolean[][] nulos,
                                 int[] clases,
                                 double[] ejemplo,
                                 double[] ejReal,
                                 int[] ejNominal,
                                 boolean[] ejNulos,
                                 int nClases,
                                 boolean distance)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulos - Reference to the training set (null values)

clases - Output attribute of each instance

ejemplo - New instance to classifiy

ejReal - New instance to classifiy (real valued)

ejNominal - New instance to classifiy (nominal valued)

ejNulos - New instance to classifiy (null values)

nClases - Number of classes of the problem

distance - True= Euclidean distance; False= HVDM

Returns:

Class of the new instance

evaluacionKNN2

public static int evaluacionKNN2(int nvec,
                                 double[][] conj,
                                 double[][] real,
                                 int[][] nominal,
                                 boolean[][] nulos,
                                 int[] clases,
                                 double[] ejemplo,
                                 double[] ejReal,
                                 int[] ejNominal,
                                 boolean[] ejNulos,
                                 int nClases,
                                 boolean distance,
                                 Referencia nVotos)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulos - Reference to the training set (null values)

clases - Output attribute of each instance

ejemplo - New instance to classifiy

ejReal - New instance to classifiy (real valued)

ejNominal - New instance to classifiy (nominal valued)

ejNulos - New instance to classifiy (null values)

nClases - Number of classes of the problem

distance - True= Euclidean distance; False= HVDM

nVotos - Maximun number of votes achieved

Returns:

Class of the new instance

evaluacionKNN3

public static int[] evaluacionKNN3(int nvec,
                                   double[][] conj,
                                   double[][] real,
                                   int[][] nominal,
                                   boolean[][] nulos,
                                   int[] clases,
                                   double[] ejemplo,
                                   double[] ejReal,
                                   int[] ejNominal,
                                   boolean[] ejNulos,
                                   int nClases,
                                   boolean distance)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulos - Reference to the training set (null values)

clases - Output attribute of each instance

ejemplo - New instance to classifiy

ejReal - New instance to classifiy (real valued)

ejNominal - New instance to classifiy (nominal valued)

ejNulos - New instance to classifiy (null values)

nClases - Number of classes of the problem

distance - True= Euclidean distance; False= HVDM

Returns:

The neighboors' classes.

evaluacionKNN2

public static int evaluacionKNN2(int nvec,
                                 double[][] conj,
                                 double[][] real,
                                 int[][] nominal,
                                 boolean[][] nulos,
                                 int[] clases,
                                 double[] ejemplo,
                                 double[] ejReal,
                                 int[] ejNominal,
                                 boolean[] ejNulos,
                                 int nClases,
                                 boolean distance,
                                 int[] vecinos)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulos - Reference to the training set (null values)

clases - Output attribute of each instance

ejemplo - New instance to classifiy

ejReal - New instance to classifiy (real valued)

ejNominal - New instance to classifiy (nominal valued)

ejNulos - New instance to classifiy (null values)

nClases - Number of classes of the problem

distance - True= Euclidean distance; False= HVDM

vecinos - Neighbors of the new instance

Returns:

Class of the new instance

evaluacionKNN2

public static int evaluacionKNN2(int nvec,
                                 double[][] conj,
                                 double[][] real,
                                 int[][] nominal,
                                 boolean[][] nulos,
                                 int[] clases,
                                 double[] ejemplo,
                                 double[] ejReal,
                                 int[] ejNominal,
                                 boolean[] ejNulos,
                                 int nClases,
                                 boolean distance,
                                 int[] vecinos,
                                 double[] minDistanciasSal,
                                 int[] votosSal)

Executes KNN

Parameters:

nvec - Number of neighbors

conj - Reference to the training set

real - Reference to the training set (real valued)

nominal - Reference to the training set (nominal valued)

nulos - Reference to the training set (null values)

clases - Output attribute of each instance

ejemplo - New instance to classifiy

ejReal - New instance to classifiy (real valued)

ejNominal - New instance to classifiy (nominal valued)

ejNulos - New instance to classifiy (null values)

nClases - Number of classes of the problem

distance - True= Euclidean distance; False= HVDM

vecinos - Neighbors of the new instance

minDistanciasSal - Resulting distances to each neighbor.

votosSal - Resulting votes for each class.

Returns:

Class of the new instance

distancia

public static double distancia(double[] ej1,
                               double[] ej1Real,
                               int[] ej1Nom,
                               boolean[] ej1Nul,
                               double[] ej2,
                               double[] ej2Real,
                               int[] ej2Nom,
                               boolean[] ej2Nul,
                               boolean Euc)

Calculates the HVDM distance between two instances

Parameters:

ej1 - First instance

ej1Real - First instance (Real valued)

ej1Nom - First instance (Nominal valued)

ej1Nul - First instance (Null values)

ej2 - Second instance

ej2Real - First instance (Real valued)

ej2Nom - First instance (Nominal valued)

ej2Nul - First instance (Null values)

Euc - Use euclidean distance instead of HVDM

Returns:

The HVDM distance

distancia2

public static double distancia2(double[] ej1,
                                double[] ej1Real,
                                int[] ej1Nom,
                                boolean[] ej1Nul,
                                double[] ej2,
                                double[] ej2Real,
                                int[] ej2Nom,
                                boolean[] ej2Nul,
                                boolean Euc)

Calculates the unsquared HVDM distance between two instances

Parameters:

ej1 - First instance

ej1Real - First instance (Real valued)

ej1Nom - First instance (Nominal valued)

ej1Nul - First instance (Null values)

ej2 - Second instance

ej2Real - First instance (Real valued)

ej2Nom - First instance (Nominal valued)

ej2Nul - First instance (Null values)

Euc - Use euclidean distance instead of HVDM

Returns:

The unsquared HVDM distance