1 | #ifndef _theplu_yat_statistics_sam_score_ |
---|
2 | #define _theplu_yat_statistics_sam_score_ |
---|
3 | |
---|
4 | // $Id: SAMScore.h 1487 2008-09-10 08:41:36Z jari $ |
---|
5 | |
---|
6 | /* |
---|
7 | Copyright (C) 2006, 2007 Jari Häkkinen, Peter Johansson |
---|
8 | Copyright (C) 2008 Peter Johansson |
---|
9 | |
---|
10 | This file is part of the yat library, http://dev.thep.lu.se/yat |
---|
11 | |
---|
12 | The yat library is free software; you can redistribute it and/or |
---|
13 | modify it under the terms of the GNU General Public License as |
---|
14 | published by the Free Software Foundation; either version 3 of the |
---|
15 | License, or (at your option) any later version. |
---|
16 | |
---|
17 | The yat library is distributed in the hope that it will be useful, |
---|
18 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
19 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
---|
20 | General Public License for more details. |
---|
21 | |
---|
22 | You should have received a copy of the GNU General Public License |
---|
23 | along with yat. If not, see <http://www.gnu.org/licenses/>. |
---|
24 | */ |
---|
25 | |
---|
26 | #include "Score.h" |
---|
27 | |
---|
28 | #include <cmath> |
---|
29 | |
---|
30 | namespace theplu { |
---|
31 | namespace yat { |
---|
32 | namespace utility { |
---|
33 | class VectorBase; |
---|
34 | } |
---|
35 | namespace classifier { |
---|
36 | class DataLookWeighted1D; |
---|
37 | } |
---|
38 | namespace statistics { |
---|
39 | |
---|
40 | /** |
---|
41 | @brief Class for score used in Significance Analysis of |
---|
42 | Microarrays (SAM). |
---|
43 | |
---|
44 | The score is similar to the Student t-test but with an added |
---|
45 | fudge factor in denominator to avoid groups with small variance |
---|
46 | getting a large score. \f$ \frac{m_x-m_y}{s+s_0} \f$ |
---|
47 | |
---|
48 | see http://www.pnas.org/cgi/content/abstract/98/9/5116 for |
---|
49 | details |
---|
50 | */ |
---|
51 | class SAMScore : public Score |
---|
52 | { |
---|
53 | |
---|
54 | public: |
---|
55 | /// |
---|
56 | /// @param s0 \f$ s_0 \f$ is a fudge factor |
---|
57 | /// @param absolute if true max(score, -score) is used |
---|
58 | /// |
---|
59 | SAMScore(const double s0, bool absolute=true); |
---|
60 | |
---|
61 | /** |
---|
62 | \f$ \frac{m_x-m_y}{s+s_0} \f$ where \f$ m = \frac{1}{n_x}\sum |
---|
63 | x_i \f$, \f$ s^2 = \left(\frac{1}{n_x}+\frac{1}{n_y} \right) |
---|
64 | \frac{\sum (x_i-m_x)^2 + \sum(y_i-m_y)^2}{n_x+n_y-2} \f$, and |
---|
65 | \f$ s_0 \f$ is the fudge factor. |
---|
66 | |
---|
67 | @return SAM score. If absolute=true absolute value of t-score |
---|
68 | is returned |
---|
69 | */ |
---|
70 | double score(const classifier::Target& target, |
---|
71 | const utility::VectorBase& value) const; |
---|
72 | |
---|
73 | /** |
---|
74 | \f$ \frac{m_x-m_y}{s+s_0} \f$ where \f$ m = \frac{\sum |
---|
75 | w_ix_i}{w_i} \f$, \f$ s_0 \f$ is the fudge factor, and \f$ s^2 |
---|
76 | = \left(\frac{1}{n_x}+\frac{1}{n_y} \right) \frac{\sum |
---|
77 | w_i(x_i-m_x)^2 + \sum w_i(y_i-m_y)^2}{n_x+n_y-2} \f$ where \f$ |
---|
78 | n \f$ is weighted version of number of data points \f$ |
---|
79 | \frac{\left(\sum w_i\right)^2}{\sum w_i^2} \f$. |
---|
80 | |
---|
81 | @return weighted version of SAM score. If absolute=true |
---|
82 | absolute value is returned |
---|
83 | */ |
---|
84 | double score(const classifier::Target& target, |
---|
85 | const classifier::DataLookupWeighted1D& value) const; |
---|
86 | |
---|
87 | /** |
---|
88 | \f$ \frac{m_x-m_y}{s+s_0} \f$ where \f$ m = \frac{\sum |
---|
89 | w_ix_i}{w_i} \f$, \f$ s_0 \f$ is the fudge factor, and \f$ s^2 |
---|
90 | = \left(\frac{1}{n_x}+\frac{1}{n_y} \right) \frac{\sum |
---|
91 | w_i(x_i-m_x)^2 + \sum w_i(y_i-m_y)^2}{n_x+n_y-2} \f$ where \f$ |
---|
92 | n \f$ is weighted version of number of data points \f$ |
---|
93 | \frac{\left(\sum w_i\right)^2}{\sum w_i^2} \f$. |
---|
94 | |
---|
95 | @return weighted version of SAM score. If absolute=true |
---|
96 | absolute value is returned |
---|
97 | */ |
---|
98 | double score(const classifier::Target& target, |
---|
99 | const utility::VectorBase& value, |
---|
100 | const utility::VectorBase& weight) const; |
---|
101 | private: |
---|
102 | double s0_; |
---|
103 | |
---|
104 | template<class T> |
---|
105 | double score(const T& positive, const T& negative) const |
---|
106 | { |
---|
107 | if(positive.n()+negative.n()<=2) |
---|
108 | return 0; |
---|
109 | double diff = positive.mean() - negative.mean(); |
---|
110 | double s2 = ( (1.0/positive.n()+1.0/negative.n()) * |
---|
111 | (positive.sum_xx_centered()+negative.sum_xx_centered()) / |
---|
112 | (positive.n()+negative.n()-2) ); |
---|
113 | if (diff<0 && absolute_) |
---|
114 | return -diff/(sqrt(s2)+s0_); |
---|
115 | return diff/(sqrt(s2)+s0_); |
---|
116 | } |
---|
117 | }; |
---|
118 | |
---|
119 | }}} // of namespace statistics, yat, and theplu |
---|
120 | |
---|
121 | #endif |
---|