• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 //===----------------------------------------------------------------------===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is dual licensed under the MIT and the University of Illinois Open
6 // Source Licenses. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // REQUIRES: long_tests
11 
12 // <random>
13 
14 // template<class IntType = int>
15 // class binomial_distribution
16 
17 // template<class _URNG> result_type operator()(_URNG& g, const param_type& parm);
18 
19 #include <random>
20 #include <numeric>
21 #include <vector>
22 #include <cassert>
23 
24 template <class T>
25 inline
26 T
sqr(T x)27 sqr(T x)
28 {
29     return x * x;
30 }
31 
main()32 int main()
33 {
34     {
35         typedef std::binomial_distribution<> D;
36         typedef D::param_type P;
37         typedef std::mt19937_64 G;
38         G g;
39         D d(16, .75);
40         P p(5, .75);
41         const int N = 1000000;
42         std::vector<D::result_type> u;
43         for (int i = 0; i < N; ++i)
44         {
45             D::result_type v = d(g, p);
46             assert(0 <= v && v <= p.t());
47             u.push_back(v);
48         }
49         double mean = std::accumulate(u.begin(), u.end(),
50                                               double(0)) / u.size();
51         double var = 0;
52         double skew = 0;
53         double kurtosis = 0;
54         for (unsigned i = 0; i < u.size(); ++i)
55         {
56             double dbl = (u[i] - mean);
57             double d2 = sqr(dbl);
58             var += d2;
59             skew += dbl * d2;
60             kurtosis += d2 * d2;
61         }
62         var /= u.size();
63         double dev = std::sqrt(var);
64         skew /= u.size() * dev * var;
65         kurtosis /= u.size() * var * var;
66         kurtosis -= 3;
67         double x_mean = p.t() * p.p();
68         double x_var = x_mean*(1-p.p());
69         double x_skew = (1-2*p.p()) / std::sqrt(x_var);
70         double x_kurtosis = (1-6*p.p()*(1-p.p())) / x_var;
71         assert(std::abs((mean - x_mean) / x_mean) < 0.01);
72         assert(std::abs((var - x_var) / x_var) < 0.01);
73         assert(std::abs((skew - x_skew) / x_skew) < 0.01);
74         assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.04);
75     }
76     {
77         typedef std::binomial_distribution<> D;
78         typedef D::param_type P;
79         typedef std::mt19937 G;
80         G g;
81         D d(16, .75);
82         P p(30, .03125);
83         const int N = 100000;
84         std::vector<D::result_type> u;
85         for (int i = 0; i < N; ++i)
86         {
87             D::result_type v = d(g, p);
88             assert(0 <= v && v <= p.t());
89             u.push_back(v);
90         }
91         double mean = std::accumulate(u.begin(), u.end(),
92                                               double(0)) / u.size();
93         double var = 0;
94         double skew = 0;
95         double kurtosis = 0;
96         for (unsigned i = 0; i < u.size(); ++i)
97         {
98             double dbl = (u[i] - mean);
99             double d2 = sqr(dbl);
100             var += d2;
101             skew += dbl * d2;
102             kurtosis += d2 * d2;
103         }
104         var /= u.size();
105         double dev = std::sqrt(var);
106         skew /= u.size() * dev * var;
107         kurtosis /= u.size() * var * var;
108         kurtosis -= 3;
109         double x_mean = p.t() * p.p();
110         double x_var = x_mean*(1-p.p());
111         double x_skew = (1-2*p.p()) / std::sqrt(x_var);
112         double x_kurtosis = (1-6*p.p()*(1-p.p())) / x_var;
113         assert(std::abs((mean - x_mean) / x_mean) < 0.01);
114         assert(std::abs((var - x_var) / x_var) < 0.01);
115         assert(std::abs((skew - x_skew) / x_skew) < 0.01);
116         assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.01);
117     }
118     {
119         typedef std::binomial_distribution<> D;
120         typedef D::param_type P;
121         typedef std::mt19937 G;
122         G g;
123         D d(16, .75);
124         P p(40, .25);
125         const int N = 1000000;
126         std::vector<D::result_type> u;
127         for (int i = 0; i < N; ++i)
128         {
129             D::result_type v = d(g, p);
130             assert(0 <= v && v <= p.t());
131             u.push_back(v);
132         }
133         double mean = std::accumulate(u.begin(), u.end(),
134                                               double(0)) / u.size();
135         double var = 0;
136         double skew = 0;
137         double kurtosis = 0;
138         for (unsigned i = 0; i < u.size(); ++i)
139         {
140             double dbl = (u[i] - mean);
141             double d2 = sqr(dbl);
142             var += d2;
143             skew += dbl * d2;
144             kurtosis += d2 * d2;
145         }
146         var /= u.size();
147         double dev = std::sqrt(var);
148         skew /= u.size() * dev * var;
149         kurtosis /= u.size() * var * var;
150         kurtosis -= 3;
151         double x_mean = p.t() * p.p();
152         double x_var = x_mean*(1-p.p());
153         double x_skew = (1-2*p.p()) / std::sqrt(x_var);
154         double x_kurtosis = (1-6*p.p()*(1-p.p())) / x_var;
155         assert(std::abs((mean - x_mean) / x_mean) < 0.01);
156         assert(std::abs((var - x_var) / x_var) < 0.01);
157         assert(std::abs((skew - x_skew) / x_skew) < 0.04);
158         assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.3);
159     }
160 }
161