Examples with CLARANS smile.clustering.CLARANS used on opensource projects

Example 1 with CLARANS

use of smile.clustering.CLARANS in project smile by haifengl.

the class CLARANSDemo method learn.

@Override
public JComponent learn() {
    try {
        numLocal = Integer.parseInt(numLocalField.getText().trim());
        if (numLocal < 5) {
            JOptionPane.showMessageDialog(this, "Toll smal NumLocal: " + numLocal, ERROR, JOptionPane.ERROR_MESSAGE);
            return null;
        }
    } catch (Exception e) {
        JOptionPane.showMessageDialog(this, "Invalid NumLocal: " + numLocalField.getText(), ERROR, JOptionPane.ERROR_MESSAGE);
        return null;
    }
    try {
        maxNeighbor = Integer.parseInt(maxNeighborField.getText().trim());
        if (maxNeighbor < 5) {
            JOptionPane.showMessageDialog(this, "Too small MaxNeighbor: " + maxNeighbor, ERROR, JOptionPane.ERROR_MESSAGE);
            return null;
        }
    } catch (Exception e) {
        JOptionPane.showMessageDialog(this, "Invalid MaxNeighbor: " + maxNeighborField.getText(), ERROR, JOptionPane.ERROR_MESSAGE);
        return null;
    }
    long clock = System.currentTimeMillis();
    CLARANS<double[]> clarans = new CLARANS<>(dataset[datasetIndex], new EuclideanDistance(), clusterNumber, maxNeighbor, numLocal);
    System.out.format("CLARANS clusterings %d samples in %dms\n", dataset[datasetIndex].length, System.currentTimeMillis() - clock);
    PlotCanvas plot = ScatterPlot.plot(clarans.medoids(), '@');
    for (int k = 0; k < clusterNumber; k++) {
        if (clarans.getClusterSize()[k] > 0) {
            double[][] cluster = new double[clarans.getClusterSize()[k]][];
            for (int i = 0, j = 0; i < dataset[datasetIndex].length; i++) {
                if (clarans.getClusterLabel()[i] == k) {
                    cluster[j++] = dataset[datasetIndex][i];
                }
            }
            plot.points(cluster, pointLegend, Palette.COLORS[k % Palette.COLORS.length]);
        }
    }
    plot.points(clarans.medoids(), '@');
    return plot;
}

Example 2 with CLARANS

use of smile.clustering.CLARANS in project smile by haifengl.

the class SmileUtils method learnGaussianRadialBasis.

/**
     * Learns Gaussian RBF function and centers from data. The centers are
     * chosen as the medoids of CLARANS. The standard deviation (i.e. width)
     * of Gaussian radial basis function is estimated as the width of each
     * cluster multiplied with a given scaling parameter r.
     * @param x the training dataset.
     * @param centers an array to store centers on output. Its length is used as k of CLARANS.
     * @param distance the distance functor.
     * @param r the scaling parameter.
     * @return Gaussian RBF functions with parameter learned from data.
     */
public static <T> GaussianRadialBasis[] learnGaussianRadialBasis(T[] x, T[] centers, Metric<T> distance, double r) {
    if (r <= 0.0) {
        throw new IllegalArgumentException("Invalid scaling parameter: " + r);
    }
    int k = centers.length;
    CLARANS<T> clarans = new CLARANS<>(x, distance, k, Math.min(100, (int) Math.round(0.01 * k * (x.length - k))));
    System.arraycopy(clarans.medoids(), 0, centers, 0, k);
    int n = x.length;
    int[] y = clarans.getClusterLabel();
    double[] sigma = new double[k];
    for (int i = 0; i < n; i++) {
        sigma[y[i]] += Math.sqr(distance.d(x[i], centers[y[i]]));
    }
    int[] ni = clarans.getClusterSize();
    GaussianRadialBasis[] rbf = new GaussianRadialBasis[k];
    for (int i = 0; i < k; i++) {
        if (ni[i] >= 5 || sigma[i] == 0.0) {
            sigma[i] = Math.sqrt(sigma[i] / ni[i]);
        } else {
            sigma[i] = Double.POSITIVE_INFINITY;
            for (int j = 0; j < k; j++) {
                if (i != j) {
                    double d = distance.d(centers[i], centers[j]);
                    if (d < sigma[i]) {
                        sigma[i] = d;
                    }
                }
            }
            sigma[i] /= 2.0;
        }
        rbf[i] = new GaussianRadialBasis(r * sigma[i]);
    }
    return rbf;
}

Example 3 with CLARANS

use of smile.clustering.CLARANS in project smile by haifengl.

the class SmileUtils method learnGaussianRadialBasis.

/**
     * Learns Gaussian RBF function and centers from data. The centers are
     * chosen as the medoids of CLARANS. The standard deviation (i.e. width)
     * of Gaussian radial basis function is estimated by the p-nearest neighbors
     * (among centers, not all samples) heuristic. A suggested value for
     * p is 2.
     * @param x the training dataset.
     * @param centers an array to store centers on output. Its length is used as k of CLARANS.
     * @param distance the distance functor.
     * @param p the number of nearest neighbors of centers to estimate the width
     * of Gaussian RBF functions.
     * @return Gaussian RBF functions with parameter learned from data.
     */
public static <T> GaussianRadialBasis[] learnGaussianRadialBasis(T[] x, T[] centers, Metric<T> distance, int p) {
    if (p < 1) {
        throw new IllegalArgumentException("Invalid number of nearest neighbors: " + p);
    }
    int k = centers.length;
    CLARANS<T> clarans = new CLARANS<>(x, distance, k, Math.min(100, (int) Math.round(0.01 * k * (x.length - k))));
    System.arraycopy(clarans.medoids(), 0, centers, 0, k);
    p = Math.min(p, k - 1);
    double[] r = new double[k];
    GaussianRadialBasis[] rbf = new GaussianRadialBasis[k];
    for (int i = 0; i < k; i++) {
        for (int j = 0; j < k; j++) {
            r[j] = distance.d(centers[i], centers[j]);
        }
        Arrays.sort(r);
        double r0 = 0.0;
        for (int j = 1; j <= p; j++) {
            r0 += r[j];
        }
        r0 /= p;
        rbf[i] = new GaussianRadialBasis(r0);
    }
    return rbf;
}