Examples of com.cloudera.cdk.morphline.shaded.org.apache.commons.math3.random.Well1024a

Package com.cloudera.cdk.morphline.shaded.org.apache.commons.math3.random

Examples of com.cloudera.cdk.morphline.shaded.org.apache.commons.math3.random.Well1024a

org.apache.commons.math3.random.Well1024a
ro.umontreal.ca/~lecuyer/myftp/papers/wellrng.pdf">Improved Long-Period Generators Based on Linear Recurrences Modulo 2 ACM Transactions on Mathematical Software, 32, 1 (2006). The errata for the paper are in wellrng-errata.txt.
@see WELL Random number generator @version $Id: Well1024a.java 1416643 2012-12-03 19:37:14Z tn $ @since 2.2


    @Test
    public void testCrossProduct() {
        // we compare accurate versus naive cross product implementations
        // on regular vectors (i.e. not extreme cases like in the previous test)
        Well1024a random = new Well1024a(885362227452043214l);
        for (int i = 0; i < 10000; ++i) {
            double ux = 10000 * random.nextDouble();
            double uy = 10000 * random.nextDouble();
            double uz = 10000 * random.nextDouble();
            double vx = 10000 * random.nextDouble();
            double vy = 10000 * random.nextDouble();
            double vz = 10000 * random.nextDouble();
            Vector3D cNaive = new Vector3D(uy * vz - uz * vy, uz * vx - ux * vz, ux * vy - uy * vx);
            Vector3D cAccurate = new Vector3D(ux, uy, uz).crossProduct(new Vector3D(vx, vy, vz));
            Assert.assertEquals(0.0, cAccurate.distance(cNaive), 6.0e-15 * cAccurate.getNorm());
        }
    }

View Full Code Here

    public void testValueWithInverseFunction() {
        final double lo = 2;
        final double hi = 3;
        final Logit f = new Logit(lo, hi);
        final Sigmoid g = new Sigmoid(lo, hi);
        RandomGenerator random = new Well1024a(0x49914cdd9f0b8db5l);
        final UnivariateDifferentiableFunction id = FunctionUtils.compose((UnivariateDifferentiableFunction) g,
                                                                (UnivariateDifferentiableFunction) f);


        for (int i = 0; i < 10; i++) {
            final double x = lo + random.nextDouble() * (hi - lo);
            Assert.assertEquals(x, id.value(new DerivativeStructure(1, 1, 0, x)).getValue(), EPS);
        }


        Assert.assertEquals(lo, id.value(new DerivativeStructure(1, 1, 0, lo)).getValue(), EPS);
        Assert.assertEquals(hi, id.value(new DerivativeStructure(1, 1, 0, hi)).getValue(), EPS);

View Full Code Here

        double[] epsilon = new double[] { 1.0e-20, 4.0e-16, 3.0e-15, 2.0e-11, 3.0e-9, 1.0e-6 };
        final double lo = 2;
        final double hi = 3;
        final Logit f = new Logit(lo, hi);
        final Sigmoid g = new Sigmoid(lo, hi);
        RandomGenerator random = new Well1024a(0x96885e9c1f81cea5l);
        final UnivariateDifferentiableFunction id =
                FunctionUtils.compose((UnivariateDifferentiableFunction) g, (UnivariateDifferentiableFunction) f);
        for (int maxOrder = 0; maxOrder < 6; ++maxOrder) {
            double max = 0;
            for (int i = 0; i < 10; i++) {
                final double x = lo + random.nextDouble() * (hi - lo);
                final DerivativeStructure dsX = new DerivativeStructure(1, maxOrder, 0, x);
                max = FastMath.max(max, FastMath.abs(dsX.getPartialDerivative(maxOrder) -
                                                     id.value(dsX).getPartialDerivative(maxOrder)));
                Assert.assertEquals(dsX.getPartialDerivative(maxOrder),
                                    id.value(dsX).getPartialDerivative(maxOrder),

View Full Code Here


    @Test
    public void testLinearCombination2() {
        // we compare accurate versus naive dot product implementations
        // on regular vectors (i.e. not extreme cases like in the previous test)
        Well1024a random = new Well1024a(553267312521321234l);


        for (int i = 0; i < 10000; ++i) {
            final double ux = 1e17 * random.nextDouble();
            final double uy = 1e17 * random.nextDouble();
            final double uz = 1e17 * random.nextDouble();
            final double vx = 1e17 * random.nextDouble();
            final double vy = 1e17 * random.nextDouble();
            final double vz = 1e17 * random.nextDouble();
            final double sInline = MathArrays.linearCombination(ux, vx,
                                                                uy, vy,
                                                                uz, vz);
            final double sArray = MathArrays.linearCombination(new double[] {ux, uy, uz},
                                                               new double[] {vx, vy, vz});

View Full Code Here

      HashSet<DoubleVector> leftDistribution, long seed) {
    List<DoubleVector> lst = new ArrayList<>(100);


    double mean1 = 250;
    double mean2 = 750;
    RandomDataImpl random = new RandomDataImpl(new Well1024a(seed));
    for (int i = 0; i < 50; i++) {
      double nextGaussian1 = random.nextGaussian(mean1, Math.sqrt(100));
      assertTrue(nextGaussian1 >= 150 && nextGaussian1 <= 350);
      double nextGaussian2 = random.nextGaussian(mean2, Math.sqrt(100));
      assertTrue(nextGaussian2 >= 650 && nextGaussian2 <= 850);

View Full Code Here

            int[] minIndices,
            int[] maxIndices,
            boolean withSymmetries,
            boolean generateNewDescriptors,
            long seed) {
        this(minDiffNDs, maxDiffNDs, minIndices, maxIndices, withSymmetries, generateNewDescriptors, new Well1024a(seed));
    }

View Full Code Here


    @Test
    public void testLinearCombination2() {
        // we compare accurate versus naive dot product implementations
        // on regular vectors (i.e. not extreme cases like in the previous test)
        Well1024a random = new Well1024a(553267312521321234l);


        for (int i = 0; i < 10000; ++i) {
            final double ux = 1e17 * random.nextDouble();
            final double uy = 1e17 * random.nextDouble();
            final double uz = 1e17 * random.nextDouble();
            final double vx = 1e17 * random.nextDouble();
            final double vy = 1e17 * random.nextDouble();
            final double vz = 1e17 * random.nextDouble();
            final double sInline = MathArrays.linearCombination(ux, vx,
                                                                uy, vy,
                                                                uz, vz);
            final double sArray = MathArrays.linearCombination(new double[] {ux, uy, uz},
                                                               new double[] {vx, vy, vz});

View Full Code Here


    @Test
    public void testDotProduct() {
        // we compare accurate versus naive dot product implementations
        // on regular vectors (i.e. not extreme cases like in the previous test)
        Well1024a random = new Well1024a(553267312521321234l);
        for (int i = 0; i < 10000; ++i) {
            double ux = 10000 * random.nextDouble();
            double uy = 10000 * random.nextDouble();
            double uz = 10000 * random.nextDouble();
            double vx = 10000 * random.nextDouble();
            double vy = 10000 * random.nextDouble();
            double vz = 10000 * random.nextDouble();
            double sNaive = ux * vx + uy * vy + uz * vz;
            double sAccurate = new Vector3D(ux, uy, uz).dotProduct(new Vector3D(vx, vy, vz));
            Assert.assertEquals(sNaive, sAccurate, 2.5e-16 * sAccurate);
        }
    }

View Full Code Here


    @Test
    public void testCrossProduct() {
        // we compare accurate versus naive cross product implementations
        // on regular vectors (i.e. not extreme cases like in the previous test)
        Well1024a random = new Well1024a(885362227452043214l);
        for (int i = 0; i < 10000; ++i) {
            double ux = 10000 * random.nextDouble();
            double uy = 10000 * random.nextDouble();
            double uz = 10000 * random.nextDouble();
            double vx = 10000 * random.nextDouble();
            double vy = 10000 * random.nextDouble();
            double vz = 10000 * random.nextDouble();
            Vector3D cNaive = new Vector3D(uy * vz - uz * vy, uz * vx - ux * vz, ux * vy - uy * vx);
            Vector3D cAccurate = new Vector3D(ux, uy, uz).crossProduct(new Vector3D(vx, vy, vz));
            Assert.assertEquals(0.0, cAccurate.distance(cNaive), 6.0e-15 * cAccurate.getNorm());
        }
    }

View Full Code Here

0 1 2 3 4 5

TOP

Related Classes of com.cloudera.cdk.morphline.shaded.org.apache.commons.math3.random.Well1024a

cc.redberry.core.tensor.random.RandomTensor

com.cloudera.cdk.morphline.stdlib.GenerateUUIDBuilder$GenerateUUID

de.jungblut.clustering.MeanShiftClusteringTest

org.apache.commons.math3.analysis.differentiation.DerivativeStructureTest

org.apache.commons.math3.analysis.differentiation.SparseGradientTest

org.apache.commons.math3.analysis.function.LogitTest

org.apache.commons.math3.ExtendedFieldElementAbstractTest

org.apache.commons.math3.geometry.enclosing.WelzlEncloser2DTest

org.apache.commons.math3.geometry.enclosing.WelzlEncloser3DTest

org.apache.commons.math3.geometry.euclidean.threed.FieldRotationDfpTest

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.