/**
* Copyright (C) 2009 - present by OpenGamma Inc. and the OpenGamma group of companies
*
* Please see distribution for license.
*/
package com.opengamma.analytics.math.statistics.distribution;
import static org.testng.AssertJUnit.assertEquals;
import static org.testng.AssertJUnit.assertFalse;
import org.testng.Assert;
import org.testng.annotations.Test;
import com.opengamma.analytics.math.function.Function1D;
import com.opengamma.analytics.math.statistics.descriptive.MeanCalculator;
import com.opengamma.analytics.math.statistics.descriptive.MedianCalculator;
import com.opengamma.analytics.math.statistics.descriptive.PopulationVarianceCalculator;
public class GeneralizedParetoDistributionTest extends ProbabilityDistributionTestCase {
private static final double MU = 0.4;
private static final double SIGMA = 1.4;
private static final double KSI = 0.2;
private static final GeneralizedParetoDistribution DIST = new GeneralizedParetoDistribution(MU, SIGMA, KSI, ENGINE);
private static final double LARGE_X = 1e20;
@Test(expectedExceptions = IllegalArgumentException.class)
public void testBadSigma() {
new GeneralizedParetoDistribution(MU, -SIGMA, KSI);
}
@Test(expectedExceptions = IllegalArgumentException.class)
public void testZeroKsi() {
new GeneralizedParetoDistribution(MU, SIGMA, 0);
}
@Test(expectedExceptions = IllegalArgumentException.class)
public void testNullEngine() {
new GeneralizedParetoDistribution(MU, SIGMA, KSI, null);
}
@Test
public void testBadInputs() {
assertCDFWithNull(DIST);
assertPDFWithNull(DIST);
}
@Test
public void testObject() {
assertEquals(KSI, DIST.getKsi(), 0);
assertEquals(MU, DIST.getMu(), 0);
assertEquals(SIGMA, DIST.getSigma(), 0);
GeneralizedParetoDistribution other = new GeneralizedParetoDistribution(MU, SIGMA, KSI, ENGINE);
assertEquals(DIST, other);
assertEquals(DIST.hashCode(), other.hashCode());
other = new GeneralizedParetoDistribution(MU, SIGMA, KSI);
assertEquals(DIST, other);
assertEquals(DIST.hashCode(), other.hashCode());
other = new GeneralizedParetoDistribution(MU + 1, SIGMA, KSI);
assertFalse(other.equals(DIST));
other = new GeneralizedParetoDistribution(MU, SIGMA + 1, KSI);
assertFalse(other.equals(DIST));
other = new GeneralizedParetoDistribution(MU, SIGMA, KSI + 1);
assertFalse(other.equals(DIST));
}
@Test
public void testSupport() {
ProbabilityDistribution<Double> dist = new GeneralizedParetoDistribution(MU, SIGMA, KSI, ENGINE);
assertLimit(dist, MU - EPS);
assertEquals(dist.getCDF(MU + EPS), 0, EPS);
assertEquals(dist.getCDF(LARGE_X), 1, EPS);
dist = new GeneralizedParetoDistribution(MU, SIGMA, -KSI);
final double limit = MU + SIGMA / KSI;
assertLimit(dist, MU - EPS);
assertLimit(dist, limit + EPS);
assertEquals(dist.getCDF(MU + EPS), 0, EPS);
assertEquals(dist.getCDF(limit - 1e-15), 1, EPS);
}
@Test
public void testDistribution() {
final Function1D<double[], Double> meanCalculator = new MeanCalculator();
final Function1D<double[], Double> medianCalculator = new MedianCalculator();
final Function1D<double[], Double> varianceCalculator = new PopulationVarianceCalculator();
final int n = 1000000;
final double eps = 0.1;
final double[] data = new double[n];
for (int i = 0; i < n; i++) {
data[i] = DIST.nextRandom();
}
final double mean = MU + SIGMA / (1 - KSI);
final double median = MU + SIGMA * (Math.pow(2, KSI) - 1) / KSI;
final double variance = SIGMA * SIGMA / ((1 - KSI) * (1 - KSI) * (1 - 2 * KSI));
assertEquals(meanCalculator.evaluate(data), mean, eps);
assertEquals(medianCalculator.evaluate(data), median, eps);
assertEquals(varianceCalculator.evaluate(data), variance, eps);
}
private void assertLimit(final ProbabilityDistribution<Double> dist, final double limit) {
try {
dist.getCDF(limit);
Assert.fail();
} catch (final IllegalArgumentException e) {
// Expected
}
try {
dist.getPDF(limit);
Assert.fail();
} catch (final IllegalArgumentException e) {
// Expected
}
}
}