Examples with GaussianKernel smile.math.kernel.GaussianKernel used on opensource projects

Example 11 with GaussianKernel

use of smile.math.kernel.GaussianKernel in project smile by haifengl.

the class GaussianProcessRegressionTest method testPuma8nh.

/**
     * Test of learn method, of class GaussianProcessRegression.
     */
@Test
public void testPuma8nh() {
    System.out.println("puma8nh");
    ArffParser parser = new ArffParser();
    parser.setResponseIndex(8);
    try {
        AttributeDataset data = parser.parse(smile.data.parser.IOUtils.getTestDataFile("weka/regression/puma8nh.arff"));
        double[] y = data.toArray(new double[data.size()]);
        double[][] x = data.toArray(new double[data.size()][]);
        int[] perm = Math.permutate(x.length);
        double[][] datax = new double[4000][];
        double[] datay = new double[datax.length];
        for (int i = 0; i < datax.length; i++) {
            datax[i] = x[perm[i]];
            datay[i] = y[perm[i]];
        }
        int n = datax.length;
        int k = 10;
        CrossValidation cv = new CrossValidation(n, k);
        double rss = 0.0;
        double sparseRSS30 = 0.0;
        for (int i = 0; i < k; i++) {
            double[][] trainx = Math.slice(datax, cv.train[i]);
            double[] trainy = Math.slice(datay, cv.train[i]);
            double[][] testx = Math.slice(datax, cv.test[i]);
            double[] testy = Math.slice(datay, cv.test[i]);
            GaussianProcessRegression<double[]> rkhs = new GaussianProcessRegression<>(trainx, trainy, new GaussianKernel(38.63), 0.1);
            KMeans kmeans = new KMeans(trainx, 30, 10);
            double[][] centers = kmeans.centroids();
            double r0 = 0.0;
            for (int l = 0; l < centers.length; l++) {
                for (int j = 0; j < l; j++) {
                    r0 += Math.distance(centers[l], centers[j]);
                }
            }
            r0 /= (2 * centers.length);
            System.out.println("Kernel width = " + r0);
            GaussianProcessRegression<double[]> sparse30 = new GaussianProcessRegression<>(trainx, trainy, centers, new GaussianKernel(r0), 0.1);
            for (int j = 0; j < testx.length; j++) {
                double r = testy[j] - rkhs.predict(testx[j]);
                rss += r * r;
                r = testy[j] - sparse30.predict(testx[j]);
                sparseRSS30 += r * r;
            }
        }
        System.out.println("Regular 10-CV MSE = " + rss / n);
        System.out.println("Sparse (30) 10-CV MSE = " + sparseRSS30 / n);
    } catch (Exception ex) {
        System.err.println(ex);
    }
}

Example 12 with GaussianKernel

use of smile.math.kernel.GaussianKernel in project smile by haifengl.

the class SVMTest method testUSPS.

/**
     * Test of learn method, of class SVM.
     */
@Test
public void testUSPS() {
    System.out.println("USPS");
    DelimitedTextParser parser = new DelimitedTextParser();
    parser.setResponseIndex(new NominalAttribute("class"), 0);
    try {
        AttributeDataset train = parser.parse("USPS Train", smile.data.parser.IOUtils.getTestDataFile("usps/zip.train"));
        AttributeDataset test = parser.parse("USPS Test", smile.data.parser.IOUtils.getTestDataFile("usps/zip.test"));
        double[][] x = train.toArray(new double[train.size()][]);
        int[] y = train.toArray(new int[train.size()]);
        double[][] testx = test.toArray(new double[test.size()][]);
        int[] testy = test.toArray(new int[test.size()]);
        SVM<double[]> svm = new SVM<>(new GaussianKernel(8.0), 5.0, Math.max(y) + 1, SVM.Multiclass.ONE_VS_ONE);
        svm.learn(x, y);
        svm.finish();
        int error = 0;
        for (int i = 0; i < testx.length; i++) {
            if (svm.predict(testx[i]) != testy[i]) {
                error++;
            }
        }
        System.out.format("USPS error rate = %.2f%%%n", 100.0 * error / testx.length);
        assertTrue(error < 95);
        System.out.println("USPS one more epoch...");
        for (int i = 0; i < x.length; i++) {
            int j = Math.randomInt(x.length);
            svm.learn(x[j], y[j]);
        }
        svm.finish();
        error = 0;
        for (int i = 0; i < testx.length; i++) {
            if (svm.predict(testx[i]) != testy[i]) {
                error++;
            }
        }
        System.out.format("USPS error rate = %.2f%%%n", 100.0 * error / testx.length);
        assertTrue(error < 95);
    } catch (Exception ex) {
        ex.printStackTrace();
    }
}

Example 13 with GaussianKernel

use of smile.math.kernel.GaussianKernel in project smile by haifengl.

the class SVMTest method testSegment.

/**
     * Test of learn method, of class SVM.
     */
@Test
public void testSegment() {
    System.out.println("Segment");
    ArffParser parser = new ArffParser();
    parser.setResponseIndex(19);
    try {
        AttributeDataset train = parser.parse(smile.data.parser.IOUtils.getTestDataFile("weka/segment-challenge.arff"));
        AttributeDataset test = parser.parse(smile.data.parser.IOUtils.getTestDataFile("weka/segment-test.arff"));
        System.out.println(train.size() + " " + test.size());
        double[][] x = train.toArray(new double[0][]);
        int[] y = train.toArray(new int[0]);
        double[][] testx = test.toArray(new double[0][]);
        int[] testy = test.toArray(new int[0]);
        SVM<double[]> svm = new SVM<>(new GaussianKernel(8.0), 5.0, Math.max(y) + 1, SVM.Multiclass.ONE_VS_ALL);
        svm.learn(x, y);
        svm.finish();
        int error = 0;
        for (int i = 0; i < testx.length; i++) {
            if (svm.predict(testx[i]) != testy[i]) {
                error++;
            }
        }
        System.out.format("Segment error rate = %.2f%%%n", 100.0 * error / testx.length);
        assertTrue(error < 70);
    } catch (Exception ex) {
        ex.printStackTrace();
    }
}

Example 14 with GaussianKernel

use of smile.math.kernel.GaussianKernel in project smile by haifengl.

the class KPCADemo method learn.

@Override
public JComponent learn() {
    JPanel pane = new JPanel(new GridLayout(2, 2));
    double[][] data = dataset[datasetIndex].toArray(new double[dataset[datasetIndex].size()][]);
    String[] names = dataset[datasetIndex].toArray(new String[dataset[datasetIndex].size()]);
    if (names[0] == null) {
        names = null;
    }
    if (gamma[datasetIndex] == 0.0) {
        int n = 0;
        for (int i = 0; i < data.length; i++) {
            for (int j = 0; j < i; j++, n++) {
                gamma[datasetIndex] += Math.squaredDistance(data[i], data[j]);
            }
        }
        gamma[datasetIndex] = Math.sqrt(gamma[datasetIndex] / n) / 4;
    } else {
        try {
            gamma[datasetIndex] = Double.parseDouble(gammaNumberField.getText().trim());
            if (gamma[datasetIndex] <= 0) {
                JOptionPane.showMessageDialog(this, "Invalid parameter: " + gamma[datasetIndex], "Error", JOptionPane.ERROR_MESSAGE);
                return null;
            }
        } catch (Exception e) {
            JOptionPane.showMessageDialog(this, "Invalid parameter: " + gammaNumberField.getText(), "Error", JOptionPane.ERROR_MESSAGE);
            return null;
        }
    }
    gammaNumberField.setText(String.format("%.4f", gamma[datasetIndex]));
    long clock = System.currentTimeMillis();
    PCA pca = new PCA(data, true);
    System.out.format("Learn PCA from %d samples in %dms\n", data.length, System.currentTimeMillis() - clock);
    pca.setProjection(2);
    double[][] y = pca.project(data);
    PlotCanvas plot = new PlotCanvas(Math.colMin(y), Math.colMax(y));
    if (names != null) {
        plot.points(y, names);
    } else if (dataset[datasetIndex].response() != null) {
        int[] labels = dataset[datasetIndex].toArray(new int[dataset[datasetIndex].size()]);
        for (int i = 0; i < y.length; i++) {
            plot.point(pointLegend, Palette.COLORS[labels[i]], y[i]);
        }
    } else {
        plot.points(y, pointLegend);
    }
    plot.setTitle("PCA");
    pane.add(plot);
    pca.setProjection(3);
    y = pca.project(data);
    plot = new PlotCanvas(Math.colMin(y), Math.colMax(y));
    if (names != null) {
        plot.points(y, names);
    } else if (dataset[datasetIndex].response() != null) {
        int[] labels = dataset[datasetIndex].toArray(new int[dataset[datasetIndex].size()]);
        for (int i = 0; i < y.length; i++) {
            plot.point(pointLegend, Palette.COLORS[labels[i]], y[i]);
        }
    } else {
        plot.points(y, pointLegend);
    }
    plot.setTitle("PCA");
    pane.add(plot);
    KPCA<double[]> kpca = new KPCA<>(data, new GaussianKernel(gamma[datasetIndex]), 2);
    y = kpca.getCoordinates();
    plot = new PlotCanvas(Math.colMin(y), Math.colMax(y));
    if (names != null) {
        plot.points(y, names);
    } else if (dataset[datasetIndex].response() != null) {
        int[] labels = dataset[datasetIndex].toArray(new int[dataset[datasetIndex].size()]);
        for (int i = 0; i < y.length; i++) {
            plot.point(pointLegend, Palette.COLORS[labels[i]], y[i]);
        }
    } else {
        plot.points(y, pointLegend);
    }
    plot.setTitle("KPCA");
    pane.add(plot);
    clock = System.currentTimeMillis();
    kpca = new KPCA<>(data, new GaussianKernel(gamma[datasetIndex]), 3);
    System.out.format("Learn KPCA from %d samples in %dms\n", data.length, System.currentTimeMillis() - clock);
    y = kpca.getCoordinates();
    plot = new PlotCanvas(Math.colMin(y), Math.colMax(y));
    if (names != null) {
        plot.points(y, names);
    } else if (dataset[datasetIndex].response() != null) {
        int[] labels = dataset[datasetIndex].toArray(new int[dataset[datasetIndex].size()]);
        for (int i = 0; i < y.length; i++) {
            plot.point(pointLegend, Palette.COLORS[labels[i]], y[i]);
        }
    } else {
        plot.points(y, pointLegend);
    }
    plot.setTitle("KPCA");
    pane.add(plot);
    return pane;
}