/*********************************************************************** This file is part of KEEL-software, the Data Mining tool for regression, classification, clustering, pattern mining and so on. Copyright (C) 2004-2010 F. Herrera (herrera@decsai.ugr.es) L. S�nchez (luciano@uniovi.es) J. Alcal�-Fdez (jalcala@decsai.ugr.es) S. Garc�a (sglopez@ujaen.es) A. Fern�ndez (alberto.fernandez@ujaen.es) J. Luengo (julianlm@decsai.ugr.es) This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/ **********************************************************************/ // // ENNTh.java // // Salvador Garc�a L�pez // // Created by Salvador Garc�a L�pez 11-7-2004. // Copyright (c) 2004 __MyCompanyName__. All rights reserved. // package keel.Algorithms.Preprocess.Instance_Selection.ENNTh; import keel.Algorithms.Preprocess.Basic.*; import org.core.*; import java.util.StringTokenizer; import java.util.Arrays; public class ENNTh extends Metodo { /*Own parameters of the algorithm*/ private int k; private double mu; public ENNTh (String ficheroScript) { super (ficheroScript); } public void ejecutar () { int i, j, l; int nClases; int claseObt; boolean marcas[]; int nSel = 0; double conjS[][]; double conjR[][]; int conjN[][]; boolean conjM[][]; int clasesS[]; int vecinos[]; double prob[]; double sumProb; double maxProb; int pos; long tiempo = System.currentTimeMillis(); /*Inicialization of the flagged instances vector for a posterior copy*/ marcas = new boolean[datosTrain.length]; for (i=0; i<datosTrain.length; i++) marcas[i] = false; /*Getting the number of differents classes*/ nClases = 0; for (i=0; i<clasesTrain.length; i++) if (clasesTrain[i] > nClases) nClases = clasesTrain[i]; nClases++; vecinos = new int[k]; prob = new double[nClases]; /*Body of the algorithm. For each instance in T, search the correspond class conform his mayority from the nearest neighborhood. Is it is positive, the instance is selected.*/ for (i=0; i<datosTrain.length; i++) { KNN.evaluacionKNN2(k, datosTrain, realTrain, nominalTrain, nulosTrain, clasesTrain, datosTrain[i], realTrain[i], nominalTrain[i], nulosTrain[i], nClases, distanceEu, vecinos); Arrays.fill(prob, 0.0); for (j=0; j<vecinos.length; j++){ if (vecinos[j]>=0) { prob[clasesTrain[vecinos[j]]] += 1.0 / (1.0 + KNN.distancia(datosTrain[i], realTrain[i], nominalTrain[i], nulosTrain[i], datosTrain[vecinos[j]], realTrain[vecinos[j]], nominalTrain[vecinos[j]], nulosTrain[vecinos[j]], distanceEu)); } } sumProb = 0.0; for (j=0; j<prob.length; j++) { sumProb += prob[j]; } for (j=0; j<prob.length; j++) { prob[j] /= sumProb; } maxProb = prob[0]; pos = 0; for (j=1; j<prob.length; j++) { if (prob[j] > maxProb) { maxProb = prob[j]; pos = j; } } claseObt = pos; if (claseObt == clasesTrain[i] && maxProb > mu) { //agree with your majority, it is included in the solution set marcas[i] = true; nSel++; } } /*Building of the S set from the flags*/ conjS = new double[nSel][datosTrain[0].length]; conjR = new double[nSel][datosTrain[0].length]; conjN = new int[nSel][datosTrain[0].length]; conjM = new boolean[nSel][datosTrain[0].length]; clasesS = new int[nSel]; for (i=0, l=0; i<datosTrain.length; i++) { if (marcas[i]) { //the instance will be copied to the solution for (j=0; j<datosTrain[0].length; j++) { conjS[l][j] = datosTrain[i][j]; conjR[l][j] = realTrain[i][j]; conjN[l][j] = nominalTrain[i][j]; conjM[l][j] = nulosTrain[i][j]; } clasesS[l] = clasesTrain[i]; l++; } } System.out.println("ENNTh "+ relation + " " + (double)(System.currentTimeMillis()-tiempo)/1000.0 + "s"); OutputIS.escribeSalida(ficheroSalida[0], conjR, conjN, conjM, clasesS, entradas, salida, nEntradas, relation); OutputIS.escribeSalida(ficheroSalida[1], test, entradas, salida, nEntradas, relation); } public void leerConfiguracion (String ficheroScript) { String fichero, linea, token; StringTokenizer lineasFichero, tokens; byte line[]; int i, j; ficheroSalida = new String[2]; fichero = Fichero.leeFichero (ficheroScript); lineasFichero = new StringTokenizer (fichero,"\n\r"); lineasFichero.nextToken(); linea = lineasFichero.nextToken(); tokens = new StringTokenizer (linea, "="); tokens.nextToken(); token = tokens.nextToken(); /*Getting the names of the training and test files*/ line = token.getBytes(); for (i=0; line[i]!='\"'; i++); i++; for (j=i; line[j]!='\"'; j++); ficheroTraining = new String (line,i,j-i); for (i=j+1; line[i]!='\"'; i++); i++; for (j=i; line[j]!='\"'; j++); ficheroTest = new String (line,i,j-i); /*Getting the path and base name of the results files*/ linea = lineasFichero.nextToken(); tokens = new StringTokenizer (linea, "="); tokens.nextToken(); token = tokens.nextToken(); /*Getting the names of output files*/ line = token.getBytes(); for (i=0; line[i]!='\"'; i++); i++; for (j=i; line[j]!='\"'; j++); ficheroSalida[0] = new String (line,i,j-i); for (i=j+1; line[i]!='\"'; i++); i++; for (j=i; line[j]!='\"'; j++); ficheroSalida[1] = new String (line,i,j-i); /*Getting the number of neighbors*/ linea = lineasFichero.nextToken(); tokens = new StringTokenizer (linea, "="); tokens.nextToken(); k = Integer.parseInt(tokens.nextToken().substring(1)); /*Getting the noise threshold*/ linea = lineasFichero.nextToken(); tokens = new StringTokenizer (linea, "="); tokens.nextToken(); mu = Double.parseDouble(tokens.nextToken().substring(1)); /*Getting the type of distance function*/ linea = lineasFichero.nextToken(); tokens = new StringTokenizer (linea, "="); tokens.nextToken(); distanceEu = tokens.nextToken().substring(1).equalsIgnoreCase("Euclidean")?true:false; } }