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 package org.apache.commons.math.distribution;
018
019 import java.io.Serializable;
020
021 import org.apache.commons.math.MathException;
022 import org.apache.commons.math.MathRuntimeException;
023 import org.apache.commons.math.exception.util.LocalizedFormats;
024 import org.apache.commons.math.special.Beta;
025 import org.apache.commons.math.util.FastMath;
026
027 /**
028 * The default implementation of {@link BinomialDistribution}.
029 *
030 * @version $Revision: 1054524 $ $Date: 2011-01-03 05:59:18 +0100 (lun. 03 janv. 2011) $
031 */
032 public class BinomialDistributionImpl extends AbstractIntegerDistribution
033 implements BinomialDistribution, Serializable {
034
035 /** Serializable version identifier */
036 private static final long serialVersionUID = 6751309484392813623L;
037
038 /** The number of trials. */
039 private int numberOfTrials;
040
041 /** The probability of success. */
042 private double probabilityOfSuccess;
043
044 /**
045 * Create a binomial distribution with the given number of trials and
046 * probability of success.
047 *
048 * @param trials the number of trials.
049 * @param p the probability of success.
050 */
051 public BinomialDistributionImpl(int trials, double p) {
052 super();
053 setNumberOfTrialsInternal(trials);
054 setProbabilityOfSuccessInternal(p);
055 }
056
057 /**
058 * Access the number of trials for this distribution.
059 *
060 * @return the number of trials.
061 */
062 public int getNumberOfTrials() {
063 return numberOfTrials;
064 }
065
066 /**
067 * Access the probability of success for this distribution.
068 *
069 * @return the probability of success.
070 */
071 public double getProbabilityOfSuccess() {
072 return probabilityOfSuccess;
073 }
074
075 /**
076 * Change the number of trials for this distribution.
077 *
078 * @param trials the new number of trials.
079 * @throws IllegalArgumentException if <code>trials</code> is not a valid
080 * number of trials.
081 * @deprecated as of 2.1 (class will become immutable in 3.0)
082 */
083 @Deprecated
084 public void setNumberOfTrials(int trials) {
085 setNumberOfTrialsInternal(trials);
086 }
087
088 /**
089 * Change the number of trials for this distribution.
090 *
091 * @param trials the new number of trials.
092 * @throws IllegalArgumentException if <code>trials</code> is not a valid
093 * number of trials.
094 */
095 private void setNumberOfTrialsInternal(int trials) {
096 if (trials < 0) {
097 throw MathRuntimeException.createIllegalArgumentException(
098 LocalizedFormats.NEGATIVE_NUMBER_OF_TRIALS, trials);
099 }
100 numberOfTrials = trials;
101 }
102
103 /**
104 * Change the probability of success for this distribution.
105 *
106 * @param p the new probability of success.
107 * @throws IllegalArgumentException if <code>p</code> is not a valid
108 * probability.
109 * @deprecated as of 2.1 (class will become immutable in 3.0)
110 */
111 @Deprecated
112 public void setProbabilityOfSuccess(double p) {
113 setProbabilityOfSuccessInternal(p);
114 }
115
116 /**
117 * Change the probability of success for this distribution.
118 *
119 * @param p the new probability of success.
120 * @throws IllegalArgumentException if <code>p</code> is not a valid
121 * probability.
122 */
123 private void setProbabilityOfSuccessInternal(double p) {
124 if (p < 0.0 || p > 1.0) {
125 throw MathRuntimeException.createIllegalArgumentException(
126 LocalizedFormats.OUT_OF_RANGE_SIMPLE, p, 0.0, 1.0);
127 }
128 probabilityOfSuccess = p;
129 }
130
131 /**
132 * Access the domain value lower bound, based on <code>p</code>, used to
133 * bracket a PDF root.
134 *
135 * @param p the desired probability for the critical value
136 * @return domain value lower bound, i.e. P(X < <i>lower bound</i>) <
137 * <code>p</code>
138 */
139 @Override
140 protected int getDomainLowerBound(double p) {
141 return -1;
142 }
143
144 /**
145 * Access the domain value upper bound, based on <code>p</code>, used to
146 * bracket a PDF root.
147 *
148 * @param p the desired probability for the critical value
149 * @return domain value upper bound, i.e. P(X < <i>upper bound</i>) >
150 * <code>p</code>
151 */
152 @Override
153 protected int getDomainUpperBound(double p) {
154 return numberOfTrials;
155 }
156
157 /**
158 * For this distribution, X, this method returns P(X ≤ x).
159 *
160 * @param x the value at which the PDF is evaluated.
161 * @return PDF for this distribution.
162 * @throws MathException if the cumulative probability can not be computed
163 * due to convergence or other numerical errors.
164 */
165 @Override
166 public double cumulativeProbability(int x) throws MathException {
167 double ret;
168 if (x < 0) {
169 ret = 0.0;
170 } else if (x >= numberOfTrials) {
171 ret = 1.0;
172 } else {
173 ret = 1.0 - Beta.regularizedBeta(getProbabilityOfSuccess(),
174 x + 1.0, numberOfTrials - x);
175 }
176 return ret;
177 }
178
179 /**
180 * For this distribution, X, this method returns P(X = x).
181 *
182 * @param x the value at which the PMF is evaluated.
183 * @return PMF for this distribution.
184 */
185 public double probability(int x) {
186 double ret;
187 if (x < 0 || x > numberOfTrials) {
188 ret = 0.0;
189 } else {
190 ret = FastMath.exp(SaddlePointExpansion.logBinomialProbability(x,
191 numberOfTrials, probabilityOfSuccess,
192 1.0 - probabilityOfSuccess));
193 }
194 return ret;
195 }
196
197 /**
198 * For this distribution, X, this method returns the largest x, such that
199 * P(X ≤ x) ≤ <code>p</code>.
200 * <p>
201 * Returns <code>-1</code> for p=0 and <code>Integer.MAX_VALUE</code> for
202 * p=1.
203 * </p>
204 *
205 * @param p the desired probability
206 * @return the largest x such that P(X ≤ x) <= p
207 * @throws MathException if the inverse cumulative probability can not be
208 * computed due to convergence or other numerical errors.
209 * @throws IllegalArgumentException if p < 0 or p > 1
210 */
211 @Override
212 public int inverseCumulativeProbability(final double p)
213 throws MathException {
214 // handle extreme values explicitly
215 if (p == 0) {
216 return -1;
217 }
218 if (p == 1) {
219 return Integer.MAX_VALUE;
220 }
221
222 // use default bisection impl
223 return super.inverseCumulativeProbability(p);
224 }
225
226 /**
227 * Returns the lower bound of the support for the distribution.
228 *
229 * The lower bound of the support is always 0 no matter the number of trials
230 * and probability parameter.
231 *
232 * @return lower bound of the support (always 0)
233 * @since 2.2
234 */
235 public int getSupportLowerBound() {
236 return 0;
237 }
238
239 /**
240 * Returns the upper bound of the support for the distribution.
241 *
242 * The upper bound of the support is the number of trials.
243 *
244 * @return upper bound of the support (equal to number of trials)
245 * @since 2.2
246 */
247 public int getSupportUpperBound() {
248 return getNumberOfTrials();
249 }
250
251 /**
252 * Returns the mean.
253 *
254 * For <code>n</code> number of trials and
255 * probability parameter <code>p</code>, the mean is
256 * <code>n * p</code>
257 *
258 * @return the mean
259 * @since 2.2
260 */
261 public double getNumericalMean() {
262 return (double)getNumberOfTrials() * getProbabilityOfSuccess();
263 }
264
265 /**
266 * Returns the variance.
267 *
268 * For <code>n</code> number of trials and
269 * probability parameter <code>p</code>, the variance is
270 * <code>n * p * (1 - p)</code>
271 *
272 * @return the variance
273 * @since 2.2
274 */
275 public double getNumericalVariance() {
276 final double p = getProbabilityOfSuccess();
277 return (double)getNumberOfTrials() * p * (1 - p);
278 }
279 }