001 /*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements. See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License. You may obtain a copy of the License at
008 *
009 * http://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 */
017
018 package org.apache.commons.math.distribution;
019
020 import org.apache.commons.math.TestUtils;
021
022 /**
023 * Test cases for HyperGeometriclDistribution.
024 * Extends IntegerDistributionAbstractTest. See class javadoc for
025 * IntegerDistributionAbstractTest for details.
026 *
027 * @version $Revision: 762087 $ $Date: 2009-04-05 10:20:18 -0400 (Sun, 05 Apr 2009) $
028 */
029 public class HypergeometricDistributionTest extends IntegerDistributionAbstractTest {
030
031 /**
032 * Constructor for ChiSquareDistributionTest.
033 * @param name
034 */
035 public HypergeometricDistributionTest(String name) {
036 super(name);
037 }
038
039 //-------------- Implementations for abstract methods -----------------------
040
041 /** Creates the default discrete distribution instance to use in tests. */
042 @Override
043 public IntegerDistribution makeDistribution() {
044 return new HypergeometricDistributionImpl(10,5, 5);
045 }
046
047 /** Creates the default probability density test input values */
048 @Override
049 public int[] makeDensityTestPoints() {
050 return new int[] {-1, 0, 1, 2, 3, 4, 5, 10};
051 }
052
053 /** Creates the default probability density test expected values */
054 @Override
055 public double[] makeDensityTestValues() {
056 return new double[] {0d, 0.003968d, 0.099206d, 0.396825d, 0.396825d,
057 0.099206d, 0.003968d, 0d};
058 }
059
060 /** Creates the default cumulative probability density test input values */
061 @Override
062 public int[] makeCumulativeTestPoints() {
063 return makeDensityTestPoints();
064 }
065
066 /** Creates the default cumulative probability density test expected values */
067 @Override
068 public double[] makeCumulativeTestValues() {
069 return new double[] {0d, .003968d, .103175d, .50000d, .896825d, .996032d,
070 1.00000d, 1d};
071 }
072
073 /** Creates the default inverse cumulative probability test input values */
074 @Override
075 public double[] makeInverseCumulativeTestPoints() {
076 return new double[] {0d, 0.001d, 0.010d, 0.025d, 0.050d, 0.100d, 0.999d,
077 0.990d, 0.975d, 0.950d, 0.900d, 1d};
078 }
079
080 /** Creates the default inverse cumulative probability density test expected values */
081 @Override
082 public int[] makeInverseCumulativeTestValues() {
083 return new int[] {-1, -1, 0, 0, 0, 0, 4, 3, 3, 3, 3, 5};
084 }
085
086 //-------------------- Additional test cases ------------------------------
087
088 /** Verify that if there are no failures, mass is concentrated on sampleSize */
089 public void testDegenerateNoFailures() throws Exception {
090 setDistribution(new HypergeometricDistributionImpl(5,5,3));
091 setCumulativeTestPoints(new int[] {-1, 0, 1, 3, 10 });
092 setCumulativeTestValues(new double[] {0d, 0d, 0d, 1d, 1d});
093 setDensityTestPoints(new int[] {-1, 0, 1, 3, 10});
094 setDensityTestValues(new double[] {0d, 0d, 0d, 1d, 0d});
095 setInverseCumulativeTestPoints(new double[] {0.1d, 0.5d});
096 setInverseCumulativeTestValues(new int[] {2, 2});
097 verifyDensities();
098 verifyCumulativeProbabilities();
099 verifyInverseCumulativeProbabilities();
100 }
101
102 /** Verify that if there are no successes, mass is concentrated on 0 */
103 public void testDegenerateNoSuccesses() throws Exception {
104 setDistribution(new HypergeometricDistributionImpl(5,0,3));
105 setCumulativeTestPoints(new int[] {-1, 0, 1, 3, 10 });
106 setCumulativeTestValues(new double[] {0d, 1d, 1d, 1d, 1d});
107 setDensityTestPoints(new int[] {-1, 0, 1, 3, 10});
108 setDensityTestValues(new double[] {0d, 1d, 0d, 0d, 0d});
109 setInverseCumulativeTestPoints(new double[] {0.1d, 0.5d});
110 setInverseCumulativeTestValues(new int[] {-1, -1});
111 verifyDensities();
112 verifyCumulativeProbabilities();
113 verifyInverseCumulativeProbabilities();
114 }
115
116 /** Verify that if sampleSize = populationSize, mass is concentrated on numberOfSuccesses */
117 public void testDegenerateFullSample() throws Exception {
118 setDistribution(new HypergeometricDistributionImpl(5,3,5));
119 setCumulativeTestPoints(new int[] {-1, 0, 1, 3, 10 });
120 setCumulativeTestValues(new double[] {0d, 0d, 0d, 1d, 1d});
121 setDensityTestPoints(new int[] {-1, 0, 1, 3, 10});
122 setDensityTestValues(new double[] {0d, 0d, 0d, 1d, 0d});
123 setInverseCumulativeTestPoints(new double[] {0.1d, 0.5d});
124 setInverseCumulativeTestValues(new int[] {2, 2});
125 verifyDensities();
126 verifyCumulativeProbabilities();
127 verifyInverseCumulativeProbabilities();
128 }
129
130 public void testPopulationSize() {
131 HypergeometricDistribution dist = new HypergeometricDistributionImpl(5,3,5);
132 try {
133 dist.setPopulationSize(-1);
134 fail("negative population size. IllegalArgumentException expected");
135 } catch(IllegalArgumentException ex) {
136 }
137
138 dist.setPopulationSize(10);
139 assertEquals(10, dist.getPopulationSize());
140 }
141
142 public void testLargeValues() {
143 int populationSize = 3456;
144 int sampleSize = 789;
145 int numberOfSucceses = 101;
146 double[][] data = {
147 {0.0, 2.75646034603961e-12, 2.75646034603961e-12, 1.0},
148 {1.0, 8.55705370142386e-11, 8.83269973602783e-11, 0.999999999997244},
149 {2.0, 1.31288129219665e-9, 1.40120828955693e-9, 0.999999999911673},
150 {3.0, 1.32724172984193e-8, 1.46736255879763e-8, 0.999999998598792},
151 {4.0, 9.94501711734089e-8, 1.14123796761385e-7, 0.999999985326375},
152 {5.0, 5.89080768883643e-7, 7.03204565645028e-7, 0.999999885876203},
153 {20.0, 0.0760051397707708, 0.27349758476299, 0.802507555007781},
154 {21.0, 0.087144222047629, 0.360641806810619, 0.72650241523701},
155 {22.0, 0.0940378846881819, 0.454679691498801, 0.639358193189381},
156 {23.0, 0.0956897500614809, 0.550369441560282, 0.545320308501199},
157 {24.0, 0.0919766921922999, 0.642346133752582, 0.449630558439718},
158 {25.0, 0.083641637261095, 0.725987771013677, 0.357653866247418},
159 {96.0, 5.93849188852098e-57, 1.0, 6.01900244560712e-57},
160 {97.0, 7.96593036832547e-59, 1.0, 8.05105570861321e-59},
161 {98.0, 8.44582921934367e-61, 1.0, 8.5125340287733e-61},
162 {99.0, 6.63604297068222e-63, 1.0, 6.670480942963e-63},
163 {100.0, 3.43501099007557e-65, 1.0, 3.4437972280786e-65},
164 {101.0, 8.78623800302957e-68, 1.0, 8.78623800302957e-68},
165 };
166
167 testHypergeometricDistributionProbabilities(populationSize, sampleSize, numberOfSucceses, data);
168 }
169
170 private void testHypergeometricDistributionProbabilities(int populationSize, int sampleSize, int numberOfSucceses, double[][] data) {
171 HypergeometricDistributionImpl dist = new HypergeometricDistributionImpl(populationSize, numberOfSucceses, sampleSize);
172 for (int i = 0; i < data.length; ++i) {
173 int x = (int)data[i][0];
174 double pdf = data[i][1];
175 double actualPdf = dist.probability(x);
176 TestUtils.assertRelativelyEquals(pdf, actualPdf, 1.0e-9);
177
178 double cdf = data[i][2];
179 double actualCdf = dist.cumulativeProbability(x);
180 TestUtils.assertRelativelyEquals(cdf, actualCdf, 1.0e-9);
181
182 double cdf1 = data[i][3];
183 double actualCdf1 = dist.upperCumulativeProbability(x);
184 TestUtils.assertRelativelyEquals(cdf1, actualCdf1, 1.0e-9);
185 }
186 }
187
188 public void testMoreLargeValues() {
189 int populationSize = 26896;
190 int sampleSize = 895;
191 int numberOfSucceses = 55;
192 double[][] data = {
193 {0.0, 0.155168304750504, 0.155168304750504, 1.0},
194 {1.0, 0.29437545000746, 0.449543754757964, 0.844831695249496},
195 {2.0, 0.273841321577003, 0.723385076334967, 0.550456245242036},
196 {3.0, 0.166488572570786, 0.889873648905753, 0.276614923665033},
197 {4.0, 0.0743969744713231, 0.964270623377076, 0.110126351094247},
198 {5.0, 0.0260542785784855, 0.990324901955562, 0.0357293766229237},
199 {20.0, 3.57101101678792e-16, 1.0, 3.78252101622096e-16},
200 {21.0, 2.00551638598312e-17, 1.0, 2.11509999433041e-17},
201 {22.0, 1.04317070180562e-18, 1.0, 1.09583608347287e-18},
202 {23.0, 5.03153504903308e-20, 1.0, 5.266538166725e-20},
203 {24.0, 2.2525984149695e-21, 1.0, 2.35003117691919e-21},
204 {25.0, 9.3677424515947e-23, 1.0, 9.74327619496943e-23},
205 {50.0, 9.83633962945521e-69, 1.0, 9.8677629437617e-69},
206 {51.0, 3.13448949497553e-71, 1.0, 3.14233143064882e-71},
207 {52.0, 7.82755221928122e-74, 1.0, 7.84193567329055e-74},
208 {53.0, 1.43662126065532e-76, 1.0, 1.43834540093295e-76},
209 {54.0, 1.72312692517348e-79, 1.0, 1.7241402776278e-79},
210 {55.0, 1.01335245432581e-82, 1.0, 1.01335245432581e-82},
211 };
212 testHypergeometricDistributionProbabilities(populationSize, sampleSize, numberOfSucceses, data);
213 }
214 }