You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by br...@apache.org on 2005/08/26 07:00:08 UTC
svn commit: r240163 - in /jakarta/commons/proper/math/branches/MATH_1_1/src:
java/org/apache/commons/math/distribution/ test/org/apache/commons/math/
test/org/apache/commons/math/distribution/
Author: brentworden
Date: Thu Aug 25 22:00:02 2005
New Revision: 240163
URL: http://svn.apache.org/viewcvs?rev=240163&view=rev
Log:
PR 36215: Fixed summation problem from adding very small point probabilities.
Modified:
jakarta/commons/proper/math/branches/MATH_1_1/src/java/org/apache/commons/math/distribution/HypergeometricDistributionImpl.java
jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/TestUtils.java
jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/distribution/HypergeometricDistributionTest.java
Modified: jakarta/commons/proper/math/branches/MATH_1_1/src/java/org/apache/commons/math/distribution/HypergeometricDistributionImpl.java
URL: http://svn.apache.org/viewcvs/jakarta/commons/proper/math/branches/MATH_1_1/src/java/org/apache/commons/math/distribution/HypergeometricDistributionImpl.java?rev=240163&r1=240162&r2=240163&view=diff
==============================================================================
--- jakarta/commons/proper/math/branches/MATH_1_1/src/java/org/apache/commons/math/distribution/HypergeometricDistributionImpl.java (original)
+++ jakarta/commons/proper/math/branches/MATH_1_1/src/java/org/apache/commons/math/distribution/HypergeometricDistributionImpl.java Thu Aug 25 22:00:02 2005
@@ -81,11 +81,9 @@
if (x < domain[0]) {
ret = 0.0;
} else if(x >= domain[1]) {
- ret = 1.0;
- } else if (x - domain[0] < domain[1] - x) {
- ret = lowerCumulativeProbability(domain[0], x, n, m, k);
+ ret = 1.0;
} else {
- ret = 1.0 - upperCumulativeProbability(x + 1, domain[1], n, m, k);
+ ret = innerCumulativeProbability(domain[0], x, 1, n, m, k);
}
return ret;
@@ -179,28 +177,6 @@
}
/**
- * For this disbution, X, this method returns P(x0 ≤ X ≤ x1). This
- * probability is computed by summing the point probabilities for the values
- * x0, x0 + 1, x0 + 2, ..., x1, in that order.
- * @param x0 the inclusive, lower bound
- * @param x1 the inclusive, upper bound
- * @param n the population size.
- * @param m number of successes in the population.
- * @param k the sample size.
- * @return P(x0 ≤ X ≤ x1).
- */
- private double lowerCumulativeProbability(
- int x0, int x1, int n, int m, int k)
- {
- double ret;
- ret = 0.0;
- for (int i = x0; i <= x1; ++i) {
- ret += probability(n, m, k, i);
- }
- return ret;
- }
-
- /**
* For this disbution, X, this method returns P(X = x).
*
* @param x the value at which the PMF is evaluated.
@@ -294,36 +270,36 @@
int[] domain = getDomain(n, m, k);
if (x < domain[0]) {
ret = 1.0;
- } else if(x >= domain[1]) {
+ } else if(x > domain[1]) {
ret = 0.0;
- } else if (x - domain[0] < domain[1] - x) {
- ret = 1.0 - lowerCumulativeProbability(domain[0], x - 1, n, m, k);
} else {
- ret = upperCumulativeProbability(x, domain[1], n, m, k);
+ ret = innerCumulativeProbability(domain[1], x, -1, n, m, k);
}
return ret;
}
-
+
/**
* For this disbution, X, this method returns P(x0 ≤ X ≤ x1). This
* probability is computed by summing the point probabilities for the values
- * x1, x1 - 1, x1 - 2, ..., x0, in that order.
+ * x0, x0 + 1, x0 + 2, ..., x1, in the order directed by dx.
* @param x0 the inclusive, lower bound
* @param x1 the inclusive, upper bound
+ * @param dx the direction of summation. 1 indicates summing from x0 to x1.
+ * 0 indicates summing from x1 to x0.
* @param n the population size.
* @param m number of successes in the population.
* @param k the sample size.
* @return P(x0 ≤ X ≤ x1).
*/
- private double upperCumulativeProbability(
- int x0, int x1, int n, int m, int k)
+ private double innerCumulativeProbability(
+ int x0, int x1, int dx, int n, int m, int k)
{
- double ret = 0.0;
- for (int i = x1; i >= x0; --i) {
- ret += probability(n, m, k, i);
+ double ret = probability(n, m, k, x0);
+ while (x0 != x1) {
+ x0 += dx;
+ ret += probability(n, m, k, x0);
}
return ret;
}
-
}
Modified: jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/TestUtils.java
URL: http://svn.apache.org/viewcvs/jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/TestUtils.java?rev=240163&r1=240162&r2=240163&view=diff
==============================================================================
--- jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/TestUtils.java (original)
+++ jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/TestUtils.java Thu Aug 25 22:00:02 2005
@@ -127,4 +127,23 @@
Assert.assertEquals("Equals check", object, object2);
Assert.assertEquals("HashCode check", object.hashCode(), object2.hashCode());
}
+
+ public static void assertRelativelyEquals(double expected, double actual, double relativeError) {
+ assertRelativelyEquals(null, expected, actual, relativeError);
+ }
+
+ public static void assertRelativelyEquals(String msg, double expected, double actual, double relativeError) {
+ if (Double.isNaN(expected)) {
+ Assert.assertTrue(msg, Double.isNaN(actual));
+ } else if (Double.isNaN(actual)) {
+ Assert.assertTrue(msg, Double.isNaN(expected));
+ } else if (Double.isInfinite(actual) || Double.isInfinite(expected)) {
+ Assert.assertEquals(expected, actual, relativeError);
+ } else if (expected == 0.0) {
+ Assert.assertEquals(msg, actual, expected, relativeError);
+ } else {
+ double x = Math.abs((expected - actual) / expected);
+ Assert.assertEquals(msg, 0.0, x, relativeError);
+ }
+ }
}
Modified: jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/distribution/HypergeometricDistributionTest.java
URL: http://svn.apache.org/viewcvs/jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/distribution/HypergeometricDistributionTest.java?rev=240163&r1=240162&r2=240163&view=diff
==============================================================================
--- jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/distribution/HypergeometricDistributionTest.java (original)
+++ jakarta/commons/proper/math/branches/MATH_1_1/src/test/org/apache/commons/math/distribution/HypergeometricDistributionTest.java Thu Aug 25 22:00:02 2005
@@ -16,6 +16,8 @@
package org.apache.commons.math.distribution;
+import org.apache.commons.math.TestUtils;
+
/**
* Test cases for HyperGeometriclDistribution.
* Extends IntegerDistributionAbstractTest. See class javadoc for
@@ -127,5 +129,78 @@
dist.setPopulationSize(10);
assertEquals(10, dist.getPopulationSize());
+ }
+
+ public void testLargeValues() {
+ int populationSize = 3456;
+ int sampleSize = 789;
+ int numberOfSucceses = 101;
+ double[][] data = {
+ {0.0, 2.75646034603961e-12, 2.75646034603961e-12, 1.0},
+ {1.0, 8.55705370142386e-11, 8.83269973602783e-11, 0.999999999997244},
+ {2.0, 1.31288129219665e-9, 1.40120828955693e-9, 0.999999999911673},
+ {3.0, 1.32724172984193e-8, 1.46736255879763e-8, 0.999999998598792},
+ {4.0, 9.94501711734089e-8, 1.14123796761385e-7, 0.999999985326375},
+ {5.0, 5.89080768883643e-7, 7.03204565645028e-7, 0.999999885876203},
+ {20.0, 0.0760051397707708, 0.27349758476299, 0.802507555007781},
+ {21.0, 0.087144222047629, 0.360641806810619, 0.72650241523701},
+ {22.0, 0.0940378846881819, 0.454679691498801, 0.639358193189381},
+ {23.0, 0.0956897500614809, 0.550369441560282, 0.545320308501199},
+ {24.0, 0.0919766921922999, 0.642346133752582, 0.449630558439718},
+ {25.0, 0.083641637261095, 0.725987771013677, 0.357653866247418},
+ {96.0, 5.93849188852098e-57, 1.0, 6.01900244560712e-57},
+ {97.0, 7.96593036832547e-59, 1.0, 8.05105570861321e-59},
+ {98.0, 8.44582921934367e-61, 1.0, 8.5125340287733e-61},
+ {99.0, 6.63604297068222e-63, 1.0, 6.670480942963e-63},
+ {100.0, 3.43501099007557e-65, 1.0, 3.4437972280786e-65},
+ {101.0, 8.78623800302957e-68, 1.0, 8.78623800302957e-68},
+ };
+
+ testHypergeometricDistributionProbabilities(populationSize, sampleSize, numberOfSucceses, data);
+ }
+
+ private void testHypergeometricDistributionProbabilities(int populationSize, int sampleSize, int numberOfSucceses, double[][] data) {
+ HypergeometricDistributionImpl dist = new HypergeometricDistributionImpl(populationSize, numberOfSucceses, sampleSize);
+ for (int i = 0; i < data.length; ++i) {
+ int x = (int)data[i][0];
+ double pdf = data[i][1];
+ double actualPdf = dist.probability(x);
+ TestUtils.assertRelativelyEquals(pdf, actualPdf, 1.0e-9);
+
+ double cdf = data[i][2];
+ double actualCdf = dist.cumulativeProbability(x);
+ TestUtils.assertRelativelyEquals(cdf, actualCdf, 1.0e-9);
+
+ double cdf1 = data[i][3];
+ double actualCdf1 = dist.upperCumulativeProbability(x);
+ TestUtils.assertRelativelyEquals(cdf1, actualCdf1, 1.0e-9);
+ }
+ }
+
+ public void testMoreLargeValues() {
+ int populationSize = 26896;
+ int sampleSize = 895;
+ int numberOfSucceses = 55;
+ double[][] data = {
+ {0.0, 0.155168304750504, 0.155168304750504, 1.0},
+ {1.0, 0.29437545000746, 0.449543754757964, 0.844831695249496},
+ {2.0, 0.273841321577003, 0.723385076334967, 0.550456245242036},
+ {3.0, 0.166488572570786, 0.889873648905753, 0.276614923665033},
+ {4.0, 0.0743969744713231, 0.964270623377076, 0.110126351094247},
+ {5.0, 0.0260542785784855, 0.990324901955562, 0.0357293766229237},
+ {20.0, 3.57101101678792e-16, 1.0, 3.78252101622096e-16},
+ {21.0, 2.00551638598312e-17, 1.0, 2.11509999433041e-17},
+ {22.0, 1.04317070180562e-18, 1.0, 1.09583608347287e-18},
+ {23.0, 5.03153504903308e-20, 1.0, 5.266538166725e-20},
+ {24.0, 2.2525984149695e-21, 1.0, 2.35003117691919e-21},
+ {25.0, 9.3677424515947e-23, 1.0, 9.74327619496943e-23},
+ {50.0, 9.83633962945521e-69, 1.0, 9.8677629437617e-69},
+ {51.0, 3.13448949497553e-71, 1.0, 3.14233143064882e-71},
+ {52.0, 7.82755221928122e-74, 1.0, 7.84193567329055e-74},
+ {53.0, 1.43662126065532e-76, 1.0, 1.43834540093295e-76},
+ {54.0, 1.72312692517348e-79, 1.0, 1.7241402776278e-79},
+ {55.0, 1.01335245432581e-82, 1.0, 1.01335245432581e-82},
+ };
+ testHypergeometricDistributionProbabilities(populationSize, sampleSize, numberOfSucceses, data);
}
}
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org