libcxx/test/std/numerics/rand/rand.dist/rand.dist.bern/rand.dist.bern.negbin/eval_param.pass.cpp

   1 //===----------------------------------------------------------------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 // REQUIRES: long_tests
  10
  11 // <random>
  12
  13 // template<class IntType = int>
  14 // class negative_binomial_distribution
  15
  16 // template<class _URNG> result_type operator()(_URNG& g, const param_type& parm);
  17
  18 #include <random>
  19 #include <cassert>
  20 #include <cmath>
  21 #include <numeric>
  22 #include <vector>
  23
  24 #include "test_macros.h"
  25
  26 template <class T>
  27 inline
  28 T
  29 sqr(T x)
  30 {
  31     return x * x;
  32 }
  33
  34 int main(int, char**)
  35 {
  36     {
  37         typedef std::negative_binomial_distribution<> D;
  38         typedef D::param_type P;
  39         typedef std::minstd_rand G;
  40         G g;
  41         D d(16, .75);
  42         P p(5, .75);
  43         const int N = 1000000;
  44         std::vector<D::result_type> u;
  45         for (int i = 0; i < N; ++i)
  46         {
  47             D::result_type v = d(g, p);
  48             assert(d.min() <= v && v <= d.max());
  49             u.push_back(v);
  50         }
  51         double mean = std::accumulate(u.begin(), u.end(),
  52                                               double(0)) / u.size();
  53         double var = 0;
  54         double skew = 0;
  55         double kurtosis = 0;
  56         for (unsigned i = 0; i < u.size(); ++i)
  57         {
  58             double dbl = (u[i] - mean);
  59             double d2 = sqr(dbl);
  60             var += d2;
  61             skew += dbl * d2;
  62             kurtosis += d2 * d2;
  63         }
  64         var /= u.size();
  65         double dev = std::sqrt(var);
  66         skew /= u.size() * dev * var;
  67         kurtosis /= u.size() * var * var;
  68         kurtosis -= 3;
  69         double x_mean = p.k() * (1 - p.p()) / p.p();
  70         double x_var = x_mean / p.p();
  71         double x_skew = (2 - p.p()) / std::sqrt(p.k() * (1 - p.p()));
  72         double x_kurtosis = 6. / p.k() + sqr(p.p()) / (p.k() * (1 - p.p()));
  73         assert(std::abs((mean - x_mean) / x_mean) < 0.01);
  74         assert(std::abs((var - x_var) / x_var) < 0.01);
  75         assert(std::abs((skew - x_skew) / x_skew) < 0.01);
  76         assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.03);
  77     }
  78     {
  79         typedef std::negative_binomial_distribution<> D;
  80         typedef D::param_type P;
  81         typedef std::mt19937 G;
  82         G g;
  83         D d(16, .75);
  84         P p(30, .03125);
  85         const int N = 1000000;
  86         std::vector<D::result_type> u;
  87         for (int i = 0; i < N; ++i)
  88         {
  89             D::result_type v = d(g, p);
  90             assert(d.min() <= v && v <= d.max());
  91             u.push_back(v);
  92         }
  93         double mean = std::accumulate(u.begin(), u.end(),
  94                                               double(0)) / u.size();
  95         double var = 0;
  96         double skew = 0;
  97         double kurtosis = 0;
  98         for (unsigned i = 0; i < u.size(); ++i)
  99         {
 100             double dbl = (u[i] - mean);
 101             double d2 = sqr(dbl);
 102             var += d2;
 103             skew += dbl * d2;
 104             kurtosis += d2 * d2;
 105         }
 106         var /= u.size();
 107         double dev = std::sqrt(var);
 108         skew /= u.size() * dev * var;
 109         kurtosis /= u.size() * var * var;
 110         kurtosis -= 3;
 111         double x_mean = p.k() * (1 - p.p()) / p.p();
 112         double x_var = x_mean / p.p();
 113         double x_skew = (2 - p.p()) / std::sqrt(p.k() * (1 - p.p()));
 114         double x_kurtosis = 6. / p.k() + sqr(p.p()) / (p.k() * (1 - p.p()));
 115         assert(std::abs((mean - x_mean) / x_mean) < 0.01);
 116         assert(std::abs((var - x_var) / x_var) < 0.01);
 117         assert(std::abs((skew - x_skew) / x_skew) < 0.02);
 118         assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.1);
 119     }
 120     {
 121         typedef std::negative_binomial_distribution<> D;
 122         typedef D::param_type P;
 123         typedef std::mt19937 G;
 124         G g;
 125         D d(16, .75);
 126         P p(40, .25);
 127         const int N = 1000000;
 128         std::vector<D::result_type> u;
 129         for (int i = 0; i < N; ++i)
 130         {
 131             D::result_type v = d(g, p);
 132             assert(d.min() <= v && v <= d.max());
 133             u.push_back(v);
 134         }
 135         double mean = std::accumulate(u.begin(), u.end(),
 136                                               double(0)) / u.size();
 137         double var = 0;
 138         double skew = 0;
 139         double kurtosis = 0;
 140         for (unsigned i = 0; i < u.size(); ++i)
 141         {
 142             double dbl = (u[i] - mean);
 143             double d2 = sqr(dbl);
 144             var += d2;
 145             skew += dbl * d2;
 146             kurtosis += d2 * d2;
 147         }
 148         var /= u.size();
 149         double dev = std::sqrt(var);
 150         skew /= u.size() * dev * var;
 151         kurtosis /= u.size() * var * var;
 152         kurtosis -= 3;
 153         double x_mean = p.k() * (1 - p.p()) / p.p();
 154         double x_var = x_mean / p.p();
 155         double x_skew = (2 - p.p()) / std::sqrt(p.k() * (1 - p.p()));
 156         double x_kurtosis = 6. / p.k() + sqr(p.p()) / (p.k() * (1 - p.p()));
 157         assert(std::abs((mean - x_mean) / x_mean) < 0.01);
 158         assert(std::abs((var - x_var) / x_var) < 0.01);
 159         assert(std::abs((skew - x_skew) / x_skew) < 0.02);
 160         assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.08);
 161     }
 162
 163   return 0;
 164 }