1 | // $Id: KolmogorovSmirnov.cc 1437 2008-08-25 17:55:00Z peter $ |
---|
2 | |
---|
3 | /* |
---|
4 | Copyright (C) 2008 Peter Johansson |
---|
5 | |
---|
6 | This file is part of the yat library, http://dev.thep.lu.se/yat |
---|
7 | |
---|
8 | The yat library is free software; you can redistribute it and/or |
---|
9 | modify it under the terms of the GNU General Public License as |
---|
10 | published by the Free Software Foundation; either version 2 of the |
---|
11 | License, or (at your option) any later version. |
---|
12 | |
---|
13 | The yat library is distributed in the hope that it will be useful, |
---|
14 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
15 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
---|
16 | General Public License for more details. |
---|
17 | |
---|
18 | You should have received a copy of the GNU General Public License |
---|
19 | along with this program; if not, write to the Free Software |
---|
20 | Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA |
---|
21 | 02111-1307, USA. |
---|
22 | */ |
---|
23 | |
---|
24 | #include "KolmogorovSmirnov.h" |
---|
25 | #include "yat/random/random.h" |
---|
26 | #include "yat/utility/stl_utility.h" |
---|
27 | |
---|
28 | #include <algorithm> |
---|
29 | #include <cassert> |
---|
30 | #include <cmath> |
---|
31 | #include <deque> |
---|
32 | #include <functional> |
---|
33 | |
---|
34 | |
---|
35 | namespace theplu { |
---|
36 | namespace yat { |
---|
37 | namespace statistics { |
---|
38 | |
---|
39 | |
---|
40 | KolmogorovSmirnov::KolmogorovSmirnov(void) |
---|
41 | : cached_(true), score_(0.0), sum_w1_(0.0), sum_w2_(0.0) |
---|
42 | { |
---|
43 | } |
---|
44 | |
---|
45 | |
---|
46 | void KolmogorovSmirnov::add(double x, bool target, double w) |
---|
47 | { |
---|
48 | if (w==0) // ignoring zero weight data |
---|
49 | return; |
---|
50 | |
---|
51 | data_.insert(std::make_pair(std::make_pair(x,w),target)); |
---|
52 | if (target){ |
---|
53 | ++n1_; |
---|
54 | sum_w1_+=w; |
---|
55 | } |
---|
56 | else { |
---|
57 | ++n2_; |
---|
58 | sum_w2_+=w; |
---|
59 | } |
---|
60 | cached_=false; |
---|
61 | } |
---|
62 | |
---|
63 | |
---|
64 | double KolmogorovSmirnov::p_value(size_t perm) const |
---|
65 | { |
---|
66 | size_t count=0; |
---|
67 | std::deque<bool> targets; |
---|
68 | typedef data_w::const_iterator const_iter; |
---|
69 | for (const_iter i(data_.begin()); i!=data_.end(); ++i) |
---|
70 | targets.push_back(i->second); |
---|
71 | |
---|
72 | for (size_t i=0; i<perm; ++i){ |
---|
73 | random::random_shuffle(targets.begin(), targets.end()); |
---|
74 | KolmogorovSmirnov ks; |
---|
75 | std::deque<bool>::const_iterator target_i(targets.begin()); |
---|
76 | const_iter end(data_.end()); |
---|
77 | for (const_iter i(data_.begin()); i!=end; ++i, ++target_i) |
---|
78 | ks.add(i->first.first, *target_i, i->first.second); |
---|
79 | |
---|
80 | if (ks.score()>=score()) |
---|
81 | ++count; |
---|
82 | } |
---|
83 | return static_cast<double>(count)/perm; |
---|
84 | } |
---|
85 | |
---|
86 | |
---|
87 | double KolmogorovSmirnov::score(void) const |
---|
88 | { |
---|
89 | if (cached_) |
---|
90 | return score_; |
---|
91 | if (!sum_w1_ || !sum_w2_) |
---|
92 | return 0.0; |
---|
93 | score_=0; |
---|
94 | std::vector<double> v; |
---|
95 | scores(v); |
---|
96 | std::less<double> l; |
---|
97 | utility::abs<double> a; |
---|
98 | using utility::make_compose_f_gx_hy; |
---|
99 | score_ = std::abs(*std::max_element(v.begin(), v.end(), |
---|
100 | make_compose_f_gx_hy(l, a, a))); |
---|
101 | return score_; |
---|
102 | } |
---|
103 | |
---|
104 | |
---|
105 | void KolmogorovSmirnov::scores(std::vector<double>& res) const |
---|
106 | { |
---|
107 | res.clear(); |
---|
108 | res.reserve(data_.size()); |
---|
109 | data_w::const_iterator iter(data_.begin()); |
---|
110 | double f1=0; |
---|
111 | double f2=0; |
---|
112 | // Peter, we should take care of ties! |
---|
113 | while(iter!=data_.end()){ |
---|
114 | if (iter->second) |
---|
115 | f1 += iter->first.second; |
---|
116 | else |
---|
117 | f2 += iter->first.second; |
---|
118 | res.push_back(f1/sum_w1_-f2/sum_w2_); |
---|
119 | ++iter; |
---|
120 | } |
---|
121 | } |
---|
122 | |
---|
123 | |
---|
124 | void KolmogorovSmirnov::reset(void) |
---|
125 | { |
---|
126 | cached_=false; |
---|
127 | sum_w1_=0; |
---|
128 | sum_w2_=0; |
---|
129 | } |
---|
130 | |
---|
131 | |
---|
132 | std::ostream& operator<<(std::ostream& os, const KolmogorovSmirnov& ks) |
---|
133 | { |
---|
134 | std::vector<double> s; |
---|
135 | ks.scores(s); |
---|
136 | for (size_t i=0; i<s.size(); ++i) |
---|
137 | os << i << "\t" << s[i] << "\n"; |
---|
138 | return os; |
---|
139 | } |
---|
140 | |
---|
141 | |
---|
142 | }}} // of namespace theplu yat statistics |
---|