Ejemplo n.º 1
0
  protected int NN(int nSel, double conj[][], double ejemplo[]) {
    double mindist, dist;
    int nneigh = -1;

    mindist = Double.POSITIVE_INFINITY;

    for (int i = 0; i < nSel; i++) {
      dist = KNN.distancia(conj[i], ejemplo);
      if (dist < mindist) {
        mindist = dist;
        nneigh = i;
      }
    }
    return nneigh;
  }
Ejemplo n.º 2
0
  public void ejecutar() {

    int i, j, l, m;
    double alfai;
    int nClases;

    int claseObt;

    boolean marcas[];
    boolean notFound;

    int init;
    int clasSel[];

    int baraje[];

    int pos, tmp;
    String instanciasIN[];
    String instanciasOUT[];

    long tiempo = System.currentTimeMillis();

    /* Getting the number of differents classes */

    nClases = 0;

    for (i = 0; i < clasesTrain.length; i++) if (clasesTrain[i] > nClases) nClases = clasesTrain[i];

    nClases++;

    /* Shuffle the train set */

    baraje = new int[datosTrain.length];

    Randomize.setSeed(semilla);

    for (i = 0; i < datosTrain.length; i++) baraje[i] = i;

    for (i = 0; i < datosTrain.length; i++) {

      pos = Randomize.Randint(i, datosTrain.length - 1);

      tmp = baraje[i];

      baraje[i] = baraje[pos];

      baraje[pos] = tmp;
    }

    /*
     * Inicialization of the flagged instaces vector for a posterior
     * elimination
     */

    marcas = new boolean[datosTrain.length];

    for (i = 0; i < datosTrain.length; i++) marcas[i] = false;

    if (datosTrain.length > 0) {

      // marcas[baraje[0]] = true; //the first instance is included always

      nSel = n_p;
      if (nSel < nClases) nSel = nClases;

    } else {

      System.err.println("Input dataset is empty");

      nSel = 0;
    }
    clasSel = new int[nClases];
    System.out.print("Selecting initial neurons... ");
    // at least, there must be 1 neuron of each class at the beginning
    init = nClases;
    for (i = 0; i < nClases && i < datosTrain.length; i++) {
      pos = Randomize.Randint(0, datosTrain.length - 1);
      tmp = 0;
      while ((clasesTrain[pos] != i || marcas[pos]) && tmp < datosTrain.length) {
        pos = (pos + 1) % datosTrain.length;
        tmp++;
      }
      if (tmp < datosTrain.length) marcas[pos] = true;
      else init--;
      // clasSel[i] = i;
    }
    for (i = init; i < Math.min(nSel, datosTrain.length); i++) {
      tmp = 0;
      pos = Randomize.Randint(0, datosTrain.length - 1);
      while (marcas[pos]) {
        pos = (pos + 1) % datosTrain.length;
        tmp++;
      }
      // if(i<nClases){
      // notFound = true;
      // do{
      // for(j=i-1;j>=0 && notFound;j--){
      // if(clasSel[j] == clasesTrain[pos])
      // notFound = false;
      // }
      // if(!notFound)
      // pos = Randomize.Randint (0, datosTrain.length-1);
      // }while(!notFound);
      // }
      // clasSel[i] = clasesTrain[pos];
      marcas[pos] = true;
      init++;
    }
    nSel = init;
    System.out.println("Initial neurons selected: " + nSel);

    /* Building of the S set from the flags */

    conjS = new double[nSel][datosTrain[0].length];

    clasesS = new int[nSel];

    for (m = 0, l = 0; m < datosTrain.length; m++) {

      if (marcas[m]) { // the instance must be copied to the solution

        for (j = 0; j < datosTrain[0].length; j++) {

          conjS[l][j] = datosTrain[m][j];
        }

        clasesS[l] = clasesTrain[m];

        l++;
      }
    }

    alfai = alpha;
    boolean change = true;
    /* Body of the LVQ algorithm. */

    // Train the network
    for (int it = 0; it < T && change; it++) {
      change = false;
      alpha = alfai;
      for (i = 1; i < datosTrain.length; i++) {
        // search for the nearest neuron to training instance
        pos = NN(nSel, conjS, datosTrain[baraje[i]]);
        // nearest neuron labels correctly the class of training
        // instance?

        if (clasesS[pos] != clasesTrain[baraje[i]]) { // NO - repel
          // the neuron
          for (j = 0; j < conjS[pos].length; j++) {
            conjS[pos][j] = conjS[pos][j] - alpha * (datosTrain[baraje[i]][j] - conjS[pos][j]);
          }
          change = true;
        } else { // YES - migrate the neuron towards the input vector
          for (j = 0; j < conjS[pos].length; j++) {
            conjS[pos][j] = conjS[pos][j] + alpha * (datosTrain[baraje[i]][j] - conjS[pos][j]);
          }
        }
        alpha = nu * alpha;
      }
      // Shuffle again the training partition
      baraje = new int[datosTrain.length];

      for (i = 0; i < datosTrain.length; i++) baraje[i] = i;

      for (i = 0; i < datosTrain.length; i++) {

        pos = Randomize.Randint(i, datosTrain.length - 1);

        tmp = baraje[i];

        baraje[i] = baraje[pos];

        baraje[pos] = tmp;
      }
    }
    System.out.println(
        "LVQ " + relation + " " + (double) (System.currentTimeMillis() - tiempo) / 1000.0 + "s");
    // Classify the train data set
    instanciasIN = new String[datosReferencia.length];
    instanciasOUT = new String[datosReferencia.length];
    for (i = 0; i < datosReferencia.length; i++) {
      /* Classify the instance selected in this iteration */
      Attribute a = Attributes.getOutputAttribute(0);

      int tipo = a.getType();
      claseObt = KNN.evaluacionKNN2(1, conjS, clasesS, datosReferencia[i], nClases);
      if (tipo != Attribute.NOMINAL) {
        instanciasIN[i] = new String(String.valueOf(clasesReferencia[i]));
        instanciasOUT[i] = new String(String.valueOf(claseObt));
      } else {
        instanciasIN[i] = new String(a.getNominalValue(clasesReferencia[i]));
        instanciasOUT[i] = new String(a.getNominalValue(claseObt));
      }
    }

    escribeSalida(
        ficheroSalida[0], instanciasIN, instanciasOUT, entradas, salida, nEntradas, relation);

    // Classify the test data set
    normalizarTest();
    instanciasIN = new String[datosTest.length];
    instanciasOUT = new String[datosTest.length];
    for (i = 0; i < datosTest.length; i++) {
      /* Classify the instance selected in this iteration */
      Attribute a = Attributes.getOutputAttribute(0);

      int tipo = a.getType();

      claseObt = KNN.evaluacionKNN2(1, conjS, clasesS, datosTest[i], nClases);
      if (tipo != Attribute.NOMINAL) {
        instanciasIN[i] = new String(String.valueOf(clasesTest[i]));
        instanciasOUT[i] = new String(String.valueOf(claseObt));
      } else {
        instanciasIN[i] = new String(a.getNominalValue(clasesTest[i]));
        instanciasOUT[i] = new String(a.getNominalValue(claseObt));
      }
    }

    escribeSalida(
        ficheroSalida[1], instanciasIN, instanciasOUT, entradas, salida, nEntradas, relation);

    // Print the network to a file
    printNetworkToFile(ficheroSalida[2], referencia.getHeader());
  }
Ejemplo n.º 3
0
  /**
   * The main method of the class that includes the operations of the algorithm. It includes all the
   * operations that the algorithm has and finishes when it writes the output information into
   * files.
   */
  public void run() {

    int S[];
    int i, j, l, m;
    int nPos = 0, nNeg = 0;
    int posID;
    int nClases;
    int pos;
    int baraje[];
    int tmp;
    double conjS[][];
    int clasesS[];
    int tamS = 0;
    int claseObt;
    int cont;
    int busq;
    boolean marcas[];
    int nSel;
    double conjS2[][];
    int clasesS2[];
    double minDist, dist;

    long tiempo = System.currentTimeMillis();

    /*CNN PART*/

    /*Count of number of positive and negative examples*/
    for (i = 0; i < clasesTrain.length; i++) {
      if (clasesTrain[i] == 0) nPos++;
      else nNeg++;
    }
    if (nPos > nNeg) {
      tmp = nPos;
      nPos = nNeg;
      nNeg = tmp;
      posID = 1;
    } else {
      posID = 0;
    }

    /*Inicialization of the candidates set*/
    S = new int[datosTrain.length];
    for (i = 0; i < S.length; i++) S[i] = Integer.MAX_VALUE;

    /*Inserting an element of mayority class*/
    Randomize.setSeed(semilla);
    pos = Randomize.Randint(0, clasesTrain.length - 1);
    while (clasesTrain[pos] == posID) pos = (pos + 1) % clasesTrain.length;
    S[tamS] = pos;
    tamS++;

    /*Insert all subset of minority class*/
    for (i = 0; i < clasesTrain.length; i++) {
      if (clasesTrain[i] == posID) {
        S[tamS] = i;
        tamS++;
      }
    }

    /*Algorithm body. We resort randomly the instances of T and compare with the rest of S.
    If an instance doesn´t classified correctly, it is inserted in S*/
    baraje = new int[datosTrain.length];
    for (i = 0; i < datosTrain.length; i++) baraje[i] = i;
    for (i = 0; i < datosTrain.length; i++) {
      pos = Randomize.Randint(i, clasesTrain.length - 1);
      tmp = baraje[i];
      baraje[i] = baraje[pos];
      baraje[pos] = tmp;
    }

    for (i = 0; i < datosTrain.length; i++) {
      if (clasesTrain[i] != posID) { // only for mayority class instances
        /*Construction of the S set from the previous vector S*/
        conjS = new double[tamS][datosTrain[0].length];
        clasesS = new int[tamS];
        for (j = 0; j < tamS; j++) {
          for (l = 0; l < datosTrain[0].length; l++) conjS[j][l] = datosTrain[S[j]][l];
          clasesS[j] = clasesTrain[S[j]];
        }

        /*Do KNN to the instance*/
        claseObt = KNN.evaluacionKNN(k, conjS, clasesS, datosTrain[baraje[i]], 2);
        if (claseObt != clasesTrain[baraje[i]]) { // fail in the class, it is included in S
          Arrays.sort(S);
          busq = Arrays.binarySearch(S, baraje[i]);
          if (busq < 0) {
            S[tamS] = baraje[i];
            tamS++;
          }
        }
      }
    }

    /*Construction of the S set from the previous vector S*/
    conjS = new double[tamS][datosTrain[0].length];
    clasesS = new int[tamS];
    for (j = 0; j < tamS; j++) {
      for (l = 0; l < datosTrain[0].length; l++) conjS[j][l] = datosTrain[S[j]][l];
      clasesS[j] = clasesTrain[S[j]];
    }

    /*TOMEK LINKS PART*/

    /*Inicialization of the instance flagged vector of the S set*/
    marcas = new boolean[conjS.length];
    for (i = 0; i < conjS.length; i++) {
      marcas[i] = true;
    }
    nSel = conjS.length;

    for (i = 0; i < conjS.length; i++) {
      minDist = Double.POSITIVE_INFINITY;
      pos = 0;
      for (j = 0; j < conjS.length; j++) {
        if (i != j) {
          dist = KNN.distancia(conjS[i], conjS[j]);
          if (dist < minDist) {
            minDist = dist;
            pos = j;
          }
        }
      }
      if (clasesS[i] != clasesS[pos]) {
        if (clasesS[i] != posID) {
          if (marcas[i] == true) {
            marcas[i] = false;
            nSel--;
          }
        } else {
          if (marcas[pos] == true) {
            marcas[pos] = false;
            nSel--;
          }
        }
      }
    }

    /*Construction of the S set from the flags*/
    conjS2 = new double[nSel][conjS[0].length];
    clasesS2 = new int[nSel];
    for (m = 0, l = 0; m < conjS.length; m++) {
      if (marcas[m]) { // the instance will evaluate
        for (j = 0; j < conjS[0].length; j++) {
          conjS2[l][j] = conjS[m][j];
        }
        clasesS2[l] = clasesS[m];
        l++;
      }
    }

    System.out.println(
        "CNN_TomekLinks "
            + relation
            + " "
            + (double) (System.currentTimeMillis() - tiempo) / 1000.0
            + "s");

    OutputIS.escribeSalida(
        ficheroSalida[0], conjS2, clasesS2, entradas, salida, nEntradas, relation);
    OutputIS.escribeSalida(ficheroSalida[1], test, entradas, salida, nEntradas, relation);
  }
Ejemplo n.º 4
0
  public void ejecutar() {

    int i, j, l, m, o;

    int nClases;

    int claseObt;

    boolean marcas[];

    double conjS[][];

    int clasesS[];

    int eleS[], eleT[];

    int bestAc, aciertos;

    int temp[];

    int pos, tmp;

    long tiempo = System.currentTimeMillis();

    /*Getting the number of different classes*/

    nClases = 0;

    for (i = 0; i < clasesTrain.length; i++) if (clasesTrain[i] > nClases) nClases = clasesTrain[i];

    nClases++;

    /*Inicialization of the flagged instance vector of the S set*/

    marcas = new boolean[datosTrain.length];

    for (i = 0; i < datosTrain.length; i++) marcas[i] = false;

    /*Allocate memory for the random selection*/

    m = (int) ((porcentaje * datosTrain.length) / 100.0);

    eleS = new int[m];

    eleT = new int[datosTrain.length - m];

    temp = new int[datosTrain.length];

    for (i = 0; i < datosTrain.length; i++) temp[i] = i;

    /** Random distribution of elements in each set */
    Randomize.setSeed(semilla);

    for (i = 0; i < eleS.length; i++) {

      pos = Randomize.Randint(i, datosTrain.length - 1);

      tmp = temp[i];

      temp[i] = temp[pos];

      temp[pos] = tmp;

      eleS[i] = temp[i];
    }

    for (i = 0; i < eleT.length; i++) {

      pos = Randomize.Randint(m + i, datosTrain.length - 1);

      tmp = temp[m + i];

      temp[m + i] = temp[pos];

      temp[pos] = tmp;

      eleT[i] = temp[m + i];
    }

    for (i = 0; i < eleS.length; i++) marcas[eleS[i]] = true;

    /*Building of the S set from the flags*/

    conjS = new double[m][datosTrain[0].length];

    clasesS = new int[m];

    for (o = 0, l = 0; o < datosTrain.length; o++) {

      if (marcas[o]) { // the instance will be evaluated

        for (j = 0; j < datosTrain[0].length; j++) {

          conjS[l][j] = datosTrain[o][j];
        }

        clasesS[l] = clasesTrain[o];

        l++;
      }
    }

    /*Evaluation of the S set*/

    bestAc = 0;

    for (i = 0; i < datosTrain.length; i++) {

      claseObt = KNN.evaluacionKNN2(k, conjS, clasesS, datosTrain[i], nClases);

      if (claseObt == clasesTrain[i]) // correct clasification
      bestAc++;
    }

    /*Body of the ENNRS algorithm. Change the S set in each iteration for instances
    of the T set until get a complete sustitution*/

    for (i = 0; i < n; i++) {

      /*Preparation the set to interchange*/

      for (j = 0; j < eleS.length; j++) {

        pos = Randomize.Randint(j, eleT.length - 1);

        tmp = eleT[j];

        eleT[j] = eleT[pos];

        eleT[pos] = tmp;
      }

      /*Interchange of instances*/

      for (j = 0; j < eleS.length; j++) {

        tmp = eleS[j];

        eleS[j] = eleT[j];

        eleT[j] = tmp;

        marcas[eleS[j]] = true;

        marcas[eleT[j]] = false;
      }

      /*Building of the S set from the flags*/

      for (o = 0, l = 0; o < datosTrain.length; o++) {

        if (marcas[o]) { // the instance will evaluate

          for (j = 0; j < datosTrain[0].length; j++) {

            conjS[l][j] = datosTrain[o][j];
          }

          clasesS[l] = clasesTrain[o];

          l++;
        }
      }

      /*Evaluation of the S set*/

      aciertos = 0;

      for (j = 0; j < datosTrain.length; j++) {

        claseObt = KNN.evaluacionKNN2(k, conjS, clasesS, datosTrain[j], nClases);

        if (claseObt == clasesTrain[j]) // correct clasification
        aciertos++;
      }

      if (aciertos > bestAc) { // keep S

        bestAc = aciertos;

      } else { // undo changes

        for (j = 0; j < eleS.length; j++) {

          tmp = eleS[j];

          eleS[j] = eleT[j];

          eleT[j] = tmp;

          marcas[eleS[j]] = true;

          marcas[eleT[j]] = false;
        }
      }
    }

    /*Building of the S set from the flags*/
    /*Building of the S set from the flags*/

    for (o = 0, l = 0; o < datosTrain.length; o++) {

      if (marcas[o]) { // the instance will evaluate

        for (j = 0; j < datosTrain[0].length; j++) {

          conjS[l][j] = datosTrain[o][j];
        }

        clasesS[l] = clasesTrain[o];

        l++;
      }
    }

    System.out.println(
        "ENNRS " + relation + " " + (double) (System.currentTimeMillis() - tiempo) / 1000.0 + "s");

    // COn conjS me vale.
    int trainRealClass[][];
    int trainPrediction[][];

    trainRealClass = new int[datosTrain.length][1];
    trainPrediction = new int[datosTrain.length][1];

    // Working on training
    for (i = 0; i < datosTrain.length; i++) {
      trainRealClass[i][0] = clasesTrain[i];
      trainPrediction[i][0] = KNN.evaluate(datosTrain[i], conjS, nClases, clasesS, this.k);
    }

    KNN.writeOutput(ficheroSalida[0], trainRealClass, trainPrediction, entradas, salida, relation);

    // Working on test
    int realClass[][] = new int[datosTest.length][1];
    int prediction[][] = new int[datosTest.length][1];

    // Check  time

    for (i = 0; i < realClass.length; i++) {
      realClass[i][0] = clasesTest[i];
      prediction[i][0] = KNN.evaluate(datosTest[i], conjS, nClases, clasesS, this.k);
    }

    KNN.writeOutput(ficheroSalida[1], realClass, prediction, entradas, salida, relation);
  }