1 | #ifndef _theplu_yat_statistics_pearson_correlation_ |
---|
2 | #define _theplu_yat_statistics_pearson_correlation_ |
---|
3 | |
---|
4 | // $Id: PearsonCorrelation.h 1139 2008-02-24 01:59:27Z peter $ |
---|
5 | |
---|
6 | /* |
---|
7 | Copyright (C) 2004, 2005 Peter Johansson |
---|
8 | Copyright (C) 2006 Jari Häkkinen, Markus Ringnér, Peter Johansson |
---|
9 | Copyright (C) 2007, 2008 Peter Johansson |
---|
10 | |
---|
11 | This file is part of the yat library, http://trac.thep.lu.se/yat |
---|
12 | |
---|
13 | The yat library is free software; you can redistribute it and/or |
---|
14 | modify it under the terms of the GNU General Public License as |
---|
15 | published by the Free Software Foundation; either version 2 of the |
---|
16 | License, or (at your option) any later version. |
---|
17 | |
---|
18 | The yat library is distributed in the hope that it will be useful, |
---|
19 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
20 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
---|
21 | General Public License for more details. |
---|
22 | |
---|
23 | You should have received a copy of the GNU General Public License |
---|
24 | along with this program; if not, write to the Free Software |
---|
25 | Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA |
---|
26 | 02111-1307, USA. |
---|
27 | */ |
---|
28 | |
---|
29 | #include "AveragerPair.h" |
---|
30 | #include "AveragerPairWeighted.h" |
---|
31 | #include "yat/classifier/Target.h" |
---|
32 | #include "yat/utility/iterator_traits.h" |
---|
33 | |
---|
34 | |
---|
35 | namespace theplu { |
---|
36 | namespace yat { |
---|
37 | namespace utility { |
---|
38 | class VectorBase; |
---|
39 | } |
---|
40 | namespace statistics { |
---|
41 | |
---|
42 | /// |
---|
43 | /// @brief Class for calculating Pearson correlation. |
---|
44 | /// |
---|
45 | class PearsonCorrelation |
---|
46 | { |
---|
47 | public: |
---|
48 | /// |
---|
49 | /// @brief The default constructor. |
---|
50 | /// |
---|
51 | PearsonCorrelation(void); |
---|
52 | |
---|
53 | /// |
---|
54 | /// @brief The destructor. |
---|
55 | /// |
---|
56 | virtual ~PearsonCorrelation(void); |
---|
57 | |
---|
58 | |
---|
59 | /** |
---|
60 | \f$ \frac{\vert \sum_i(x_i-\bar{x})(y_i-\bar{y})\vert |
---|
61 | }{\sqrt{\sum_i (x_i-\bar{x})^2\sum_i (x_i-\bar{x})^2}} \f$. |
---|
62 | |
---|
63 | |
---|
64 | If ForwardIterator is weighted correlation is calculated as |
---|
65 | \f$ \frac{\vert \sum_iw^2_i(x_i-\bar{x})(y_i-\bar{y})\vert } |
---|
66 | {\sqrt{\sum_iw^2_i(x_i-\bar{x})^2\sum_iw^2_i(y_i-\bar{y})^2}} |
---|
67 | \f$, where \f$ m_x = \frac{\sum w_ix_i}{\sum w_i} \f$ and \f$ |
---|
68 | m_x = \frac{\sum w_ix_i}{\sum w_i} \f$. This expression is |
---|
69 | chosen to get a correlation equal to unity when \a x and \a y |
---|
70 | are equal. |
---|
71 | |
---|
72 | @return Pearson correlation, if absolute=true absolute value |
---|
73 | of Pearson is used. |
---|
74 | */ |
---|
75 | template<typename ForwardIterator> |
---|
76 | double score(const classifier::Target& target, |
---|
77 | ForwardIterator first, ForwardIterator last); |
---|
78 | |
---|
79 | /** |
---|
80 | \f$ \frac{\vert \sum_iw^2_i(x_i-\bar{x})(y_i-\bar{y})\vert } |
---|
81 | {\sqrt{\sum_iw^2_i(x_i-\bar{x})^2\sum_iw^2_i(y_i-\bar{y})^2}} |
---|
82 | \f$, where \f$ m_x = \frac{\sum w_ix_i}{\sum w_i} \f$ and \f$ |
---|
83 | m_x = \frac{\sum w_ix_i}{\sum w_i} \f$. This expression is |
---|
84 | chosen to get a correlation equal to unity when \a x and \a y |
---|
85 | are equal. |
---|
86 | |
---|
87 | \return absolute value of weighted version of Pearson |
---|
88 | correlation. |
---|
89 | |
---|
90 | \note ietartors must be non-weighted |
---|
91 | */ |
---|
92 | template<typename ForwardIterator1, typename ForwardIterator2> |
---|
93 | double score(const classifier::Target& target, |
---|
94 | ForwardIterator1 first1, ForwardIterator1 last1, |
---|
95 | ForwardIterator2 first2); |
---|
96 | |
---|
97 | /** |
---|
98 | The p-value is the probability of getting a correlation as |
---|
99 | large (or larger) as the observed value by random chance, when the true |
---|
100 | correlation is zero (and the data is Gaussian). |
---|
101 | |
---|
102 | @note This function can only be used together with the |
---|
103 | unweighted score. |
---|
104 | |
---|
105 | @return one-sided p-value |
---|
106 | */ |
---|
107 | double p_value_one_sided() const; |
---|
108 | |
---|
109 | private: |
---|
110 | double r_; |
---|
111 | int nof_samples_; |
---|
112 | |
---|
113 | template<typename ForwardIterator> |
---|
114 | double score(const classifier::Target& target, |
---|
115 | ForwardIterator first, ForwardIterator last, |
---|
116 | utility::unweighted_iterator_tag); |
---|
117 | |
---|
118 | template<typename ForwardIterator> |
---|
119 | double score(const classifier::Target& target, |
---|
120 | ForwardIterator first, ForwardIterator last, |
---|
121 | utility::weighted_iterator_tag); |
---|
122 | |
---|
123 | }; |
---|
124 | |
---|
125 | template<typename ForwardIterator> |
---|
126 | double PearsonCorrelation::score(const classifier::Target& target, |
---|
127 | ForwardIterator first, |
---|
128 | ForwardIterator last) |
---|
129 | { |
---|
130 | nof_samples_ = target.size(); |
---|
131 | using utility::yat_assert; |
---|
132 | yat_assert<std::runtime_error>("PearsonCorrelation: sizes mismatch"); |
---|
133 | r_ = score(target, first, last, |
---|
134 | utility::iterator_traits<ForwardIterator>::type()); |
---|
135 | return r_; |
---|
136 | } |
---|
137 | |
---|
138 | |
---|
139 | template<typename ForwardIterator> |
---|
140 | double PearsonCorrelation::score(const classifier::Target& target, |
---|
141 | ForwardIterator first, |
---|
142 | ForwardIterator last, |
---|
143 | utility::unweighted_iterator_tag tag) |
---|
144 | |
---|
145 | { |
---|
146 | AveragerPair ap; |
---|
147 | for (size_t i=0; first!=last; ++first, ++i) |
---|
148 | ap.add(target.binary(i), *first); |
---|
149 | nof_samples_ = ap.n(); |
---|
150 | return ap.correlation(); |
---|
151 | } |
---|
152 | |
---|
153 | |
---|
154 | template<typename ForwardIterator> |
---|
155 | double PearsonCorrelation::score(const classifier::Target& target, |
---|
156 | ForwardIterator first, |
---|
157 | ForwardIterator last, |
---|
158 | utility::weighted_iterator_tag tag) |
---|
159 | |
---|
160 | { |
---|
161 | AveragerPairWeighted ap; |
---|
162 | for (size_t i=0; first!=last; ++first, ++i) |
---|
163 | ap.add(target.binary(i), first.data(), 1.0, first.weight()); |
---|
164 | nof_samples_ = ap.n(); |
---|
165 | return ap.correlation(); |
---|
166 | } |
---|
167 | |
---|
168 | template<typename ForwardIterator1, typename ForwardIterator2> |
---|
169 | double PearsonCorrelation::score(const classifier::Target& target, |
---|
170 | ForwardIterator1 first1, |
---|
171 | ForwardIterator1 last1, |
---|
172 | ForwardIterator2 first2) |
---|
173 | { |
---|
174 | utility::check_iterator_is_unweighted(first1); |
---|
175 | utility::check_iterator_is_unweighted(first2); |
---|
176 | AveragerPairWeighted ap; |
---|
177 | for (size_t i=0; first1!=last1; ++first1, ++i, ++first2) |
---|
178 | ap.add(target.binary(i), *first1, 1.0, *first2); |
---|
179 | nof_samples_ = ap.n(); |
---|
180 | r_ = ap.correlation(); |
---|
181 | } |
---|
182 | |
---|
183 | }}} // of namespace statistics, yat, and theplu |
---|
184 | |
---|
185 | #endif |
---|