• 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 RealType = double>
15 // class extreme_value_distribution
16 
17 // template<class _URNG> result_type operator()(_URNG& g);
18 
19 #include <random>
20 #include <cassert>
21 #include <vector>
22 #include <numeric>
23 
24 template <class T>
25 inline
26 T
sqr(T x)27 sqr(T x)
28 {
29     return x * x;
30 }
31 
32 void
test1()33 test1()
34 {
35     typedef std::extreme_value_distribution<> D;
36     typedef std::mt19937 G;
37     G g;
38     D d(0.5, 2);
39     const int N = 1000000;
40     std::vector<D::result_type> u;
41     for (int i = 0; i < N; ++i)
42     {
43         D::result_type v = d(g);
44         u.push_back(v);
45     }
46     double mean = std::accumulate(u.begin(), u.end(), 0.0) / u.size();
47     double var = 0;
48     double skew = 0;
49     double kurtosis = 0;
50     for (unsigned i = 0; i < u.size(); ++i)
51     {
52         double dbl = (u[i] - mean);
53         double d2 = sqr(dbl);
54         var += d2;
55         skew += dbl * d2;
56         kurtosis += d2 * d2;
57     }
58     var /= u.size();
59     double dev = std::sqrt(var);
60     skew /= u.size() * dev * var;
61     kurtosis /= u.size() * var * var;
62     kurtosis -= 3;
63     double x_mean = d.a() + d.b() * 0.577215665;
64     double x_var = sqr(d.b()) * 1.644934067;
65     double x_skew = 1.139547;
66     double x_kurtosis = 12./5;
67     assert(std::abs((mean - x_mean) / x_mean) < 0.01);
68     assert(std::abs((var - x_var) / x_var) < 0.01);
69     assert(std::abs((skew - x_skew) / x_skew) < 0.01);
70     assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.01);
71 }
72 
73 void
test2()74 test2()
75 {
76     typedef std::extreme_value_distribution<> D;
77     typedef std::mt19937 G;
78     G g;
79     D d(1, 2);
80     const int N = 1000000;
81     std::vector<D::result_type> u;
82     for (int i = 0; i < N; ++i)
83     {
84         D::result_type v = d(g);
85         u.push_back(v);
86     }
87     double mean = std::accumulate(u.begin(), u.end(), 0.0) / u.size();
88     double var = 0;
89     double skew = 0;
90     double kurtosis = 0;
91     for (unsigned i = 0; i < u.size(); ++i)
92     {
93         double dbl = (u[i] - mean);
94         double d2 = sqr(dbl);
95         var += d2;
96         skew += dbl * d2;
97         kurtosis += d2 * d2;
98     }
99     var /= u.size();
100     double dev = std::sqrt(var);
101     skew /= u.size() * dev * var;
102     kurtosis /= u.size() * var * var;
103     kurtosis -= 3;
104     double x_mean = d.a() + d.b() * 0.577215665;
105     double x_var = sqr(d.b()) * 1.644934067;
106     double x_skew = 1.139547;
107     double x_kurtosis = 12./5;
108     assert(std::abs((mean - x_mean) / x_mean) < 0.01);
109     assert(std::abs((var - x_var) / x_var) < 0.01);
110     assert(std::abs((skew - x_skew) / x_skew) < 0.01);
111     assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.01);
112 }
113 
114 void
test3()115 test3()
116 {
117     typedef std::extreme_value_distribution<> D;
118     typedef std::mt19937 G;
119     G g;
120     D d(1.5, 3);
121     const int N = 1000000;
122     std::vector<D::result_type> u;
123     for (int i = 0; i < N; ++i)
124     {
125         D::result_type v = d(g);
126         u.push_back(v);
127     }
128     double mean = std::accumulate(u.begin(), u.end(), 0.0) / u.size();
129     double var = 0;
130     double skew = 0;
131     double kurtosis = 0;
132     for (unsigned i = 0; i < u.size(); ++i)
133     {
134         double dbl = (u[i] - mean);
135         double d2 = sqr(dbl);
136         var += d2;
137         skew += dbl * d2;
138         kurtosis += d2 * d2;
139     }
140     var /= u.size();
141     double dev = std::sqrt(var);
142     skew /= u.size() * dev * var;
143     kurtosis /= u.size() * var * var;
144     kurtosis -= 3;
145     double x_mean = d.a() + d.b() * 0.577215665;
146     double x_var = sqr(d.b()) * 1.644934067;
147     double x_skew = 1.139547;
148     double x_kurtosis = 12./5;
149     assert(std::abs((mean - x_mean) / x_mean) < 0.01);
150     assert(std::abs((var - x_var) / x_var) < 0.01);
151     assert(std::abs((skew - x_skew) / x_skew) < 0.01);
152     assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.01);
153 }
154 
155 void
test4()156 test4()
157 {
158     typedef std::extreme_value_distribution<> D;
159     typedef std::mt19937 G;
160     G g;
161     D d(3, 4);
162     const int N = 1000000;
163     std::vector<D::result_type> u;
164     for (int i = 0; i < N; ++i)
165     {
166         D::result_type v = d(g);
167         u.push_back(v);
168     }
169     double mean = std::accumulate(u.begin(), u.end(), 0.0) / u.size();
170     double var = 0;
171     double skew = 0;
172     double kurtosis = 0;
173     for (unsigned i = 0; i < u.size(); ++i)
174     {
175         double dbl = (u[i] - mean);
176         double d2 = sqr(dbl);
177         var += d2;
178         skew += dbl * d2;
179         kurtosis += d2 * d2;
180     }
181     var /= u.size();
182     double dev = std::sqrt(var);
183     skew /= u.size() * dev * var;
184     kurtosis /= u.size() * var * var;
185     kurtosis -= 3;
186     double x_mean = d.a() + d.b() * 0.577215665;
187     double x_var = sqr(d.b()) * 1.644934067;
188     double x_skew = 1.139547;
189     double x_kurtosis = 12./5;
190     assert(std::abs((mean - x_mean) / x_mean) < 0.01);
191     assert(std::abs((var - x_var) / x_var) < 0.01);
192     assert(std::abs((skew - x_skew) / x_skew) < 0.01);
193     assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.01);
194 }
195 
main()196 int main()
197 {
198     test1();
199     test2();
200     test3();
201     test4();
202 }
203