Coverage Report - org.apache.commons.math.distribution.AbstractContinuousDistribution

Classes in this File Line Coverage Branch Coverage Complexity
AbstractContinuousDistribution
64% 
33% 
2.833

 1  
 /*
 2  
  * Copyright 2003-2004 The Apache Software Foundation.
 3  
  *
 4  
  * Licensed under the Apache License, Version 2.0 (the "License");
 5  
  * you may not use this file except in compliance with the License.
 6  
  * You may obtain a copy of the License at
 7  
  *
 8  
  *      http://www.apache.org/licenses/LICENSE-2.0
 9  
  *
 10  
  * Unless required by applicable law or agreed to in writing, software
 11  
  * distributed under the License is distributed on an "AS IS" BASIS,
 12  
  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 13  
  * See the License for the specific language governing permissions and
 14  
  * limitations under the License.
 15  
  */
 16  
 package org.apache.commons.math.distribution;
 17  
 
 18  
 import java.io.Serializable;
 19  
 
 20  
 import org.apache.commons.math.ConvergenceException;
 21  
 import org.apache.commons.math.FunctionEvaluationException;
 22  
 import org.apache.commons.math.MathException;
 23  
 import org.apache.commons.math.analysis.UnivariateRealFunction;
 24  
 import org.apache.commons.math.analysis.UnivariateRealSolverUtils;
 25  
 
 26  
 /**
 27  
  * Base class for continuous distributions.  Default implementations are
 28  
  * provided for some of the methods that do not vary from distribution to
 29  
  * distribution.
 30  
  *  
 31  
  * @version $Revision$ $Date: 2005-02-26 05:11:52 -0800 (Sat, 26 Feb 2005) $
 32  
  */
 33  
 public abstract class AbstractContinuousDistribution
 34  
     extends AbstractDistribution
 35  
     implements ContinuousDistribution, Serializable {
 36  
 
 37  
     /** Serializable version identifier */
 38  
     static final long serialVersionUID = -38038050983108802L;
 39  
     
 40  
     /**
 41  
      * Default constructor.
 42  
      */
 43  
     protected AbstractContinuousDistribution() {
 44  546
         super();
 45  546
     }
 46  
 
 47  
     /**
 48  
      * For this distribution, X, this method returns the critical point x, such
 49  
      * that P(X &lt; x) = <code>p</code>.
 50  
      *
 51  
      * @param p the desired probability
 52  
      * @return x, such that P(X &lt; x) = <code>p</code>
 53  
      * @throws MathException if the inverse cumulative probability can not be
 54  
      *         computed due to convergence or other numerical errors.
 55  
      * @throws IllegalArgumentException if <code>p</code> is not a valid
 56  
      *         probability.
 57  
      */
 58  
     public double inverseCumulativeProbability(final double p)
 59  
         throws MathException {
 60  182
         if (p < 0.0 || p > 1.0) {
 61  20
             throw new IllegalArgumentException("p must be between 0.0 and 1.0, inclusive.");
 62  
         }
 63  
 
 64  
         // by default, do simple root finding using bracketing and default solver.
 65  
         // subclasses can overide if there is a better method.
 66  162
         UnivariateRealFunction rootFindingFunction =
 67  
             new UnivariateRealFunction() {
 68  
 
 69  162
             public double value(double x) throws FunctionEvaluationException {
 70  
                 try {
 71  5494
                     return cumulativeProbability(x) - p;
 72  0
                 } catch (MathException ex) {
 73  0
                     throw new FunctionEvaluationException
 74  
                         (x, "Error computing cdf", ex);
 75  
                 }
 76  
             }
 77  
         };
 78  
               
 79  
         // Try to bracket root, test domain endoints if this fails     
 80  162
         double lowerBound = getDomainLowerBound(p);
 81  162
         double upperBound = getDomainUpperBound(p);
 82  162
         double[] bracket = null;
 83  
         try {
 84  162
             bracket = UnivariateRealSolverUtils.bracket(
 85  
                     rootFindingFunction, getInitialDomain(p),
 86  
                     lowerBound, upperBound);
 87  0
         }  catch (ConvergenceException ex) {
 88  
             /* 
 89  
              * Check domain endpoints to see if one gives value that is within
 90  
              * the default solver's defaultAbsoluteAccuracy of 0 (will be the
 91  
              * case if density has bounded support and p is 0 or 1).
 92  
              * 
 93  
              * TODO: expose the default solver, defaultAbsoluteAccuracy as
 94  
              * a constant.
 95  
              */ 
 96  0
             if (Math.abs(rootFindingFunction.value(lowerBound)) < 1E-6) {
 97  0
                 return lowerBound;
 98  
             }
 99  0
             if (Math.abs(rootFindingFunction.value(upperBound)) < 1E-6) {
 100  0
                 return upperBound;
 101  
             }     
 102  
             // Failed bracket convergence was not because of corner solution
 103  0
             throw new MathException(ex);
 104  162
         }
 105  
 
 106  
         // find root
 107  162
         double root = UnivariateRealSolverUtils.solve(rootFindingFunction,
 108  
                 bracket[0],bracket[1]);
 109  162
         return root;
 110  
     }
 111  
 
 112  
     /**
 113  
      * Access the initial domain value, based on <code>p</code>, used to
 114  
      * bracket a CDF root.  This method is used by
 115  
      * {@link #inverseCumulativeProbability(double)} to find critical values.
 116  
      * 
 117  
      * @param p the desired probability for the critical value
 118  
      * @return initial domain value
 119  
      */
 120  
     protected abstract double getInitialDomain(double p);
 121  
 
 122  
     /**
 123  
      * Access the domain value lower bound, based on <code>p</code>, used to
 124  
      * bracket a CDF root.  This method is used by
 125  
      * {@link #inverseCumulativeProbability(double)} to find critical values.
 126  
      * 
 127  
      * @param p the desired probability for the critical value
 128  
      * @return domain value lower bound, i.e.
 129  
      *         P(X &lt; <i>lower bound</i>) &lt; <code>p</code> 
 130  
      */
 131  
     protected abstract double getDomainLowerBound(double p);
 132  
 
 133  
     /**
 134  
      * Access the domain value upper bound, based on <code>p</code>, used to
 135  
      * bracket a CDF root.  This method is used by
 136  
      * {@link #inverseCumulativeProbability(double)} to find critical values.
 137  
      * 
 138  
      * @param p the desired probability for the critical value
 139  
      * @return domain value upper bound, i.e.
 140  
      *         P(X &lt; <i>upper bound</i>) &gt; <code>p</code> 
 141  
      */
 142  
     protected abstract double getDomainUpperBound(double p);
 143  
 }