1521 |
21 Sep 08 |
peter |
1 |
#ifndef _theplu_yat_normalizer_z_score_ |
1521 |
21 Sep 08 |
peter |
2 |
#define _theplu_yat_normalizer_z_score_ |
1445 |
27 Aug 08 |
peter |
3 |
|
1575 |
14 Oct 08 |
jari |
// $Id$ |
1575 |
14 Oct 08 |
jari |
5 |
|
1445 |
27 Aug 08 |
peter |
6 |
/* |
2119 |
12 Dec 09 |
peter |
Copyright (C) 2008 Jari Häkkinen, Peter Johansson |
4089 |
07 Sep 21 |
peter |
Copyright (C) 2009, 2010, 2016, 2021 Peter Johansson |
1445 |
27 Aug 08 |
peter |
9 |
|
1445 |
27 Aug 08 |
peter |
This file is part of the yat library, http://dev.thep.lu.se/yat |
1445 |
27 Aug 08 |
peter |
11 |
|
1445 |
27 Aug 08 |
peter |
The yat library is free software; you can redistribute it and/or |
1445 |
27 Aug 08 |
peter |
modify it under the terms of the GNU General Public License as |
1486 |
09 Sep 08 |
jari |
published by the Free Software Foundation; either version 3 of the |
1445 |
27 Aug 08 |
peter |
License, or (at your option) any later version. |
1445 |
27 Aug 08 |
peter |
16 |
|
1445 |
27 Aug 08 |
peter |
The yat library is distributed in the hope that it will be useful, |
1445 |
27 Aug 08 |
peter |
but WITHOUT ANY WARRANTY; without even the implied warranty of |
1445 |
27 Aug 08 |
peter |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
1445 |
27 Aug 08 |
peter |
General Public License for more details. |
1445 |
27 Aug 08 |
peter |
21 |
|
1445 |
27 Aug 08 |
peter |
You should have received a copy of the GNU General Public License |
1487 |
10 Sep 08 |
jari |
along with yat. If not, see <http://www.gnu.org/licenses/>. |
1445 |
27 Aug 08 |
peter |
24 |
*/ |
1445 |
27 Aug 08 |
peter |
25 |
|
2157 |
18 Jan 10 |
peter |
26 |
#include "utility.h" |
2157 |
18 Jan 10 |
peter |
27 |
|
1521 |
21 Sep 08 |
peter |
28 |
#include "yat/statistics/Averager.h" |
1524 |
23 Sep 08 |
peter |
29 |
#include "yat/statistics/AveragerWeighted.h" |
1445 |
27 Aug 08 |
peter |
30 |
|
3541 |
23 Dec 16 |
peter |
31 |
#include "yat/utility/concept_check.h" |
1524 |
23 Sep 08 |
peter |
32 |
#include "yat/utility/iterator_traits.h" |
1524 |
23 Sep 08 |
peter |
33 |
|
2157 |
18 Jan 10 |
peter |
34 |
#include <boost/concept_check.hpp> |
2157 |
18 Jan 10 |
peter |
35 |
|
1445 |
27 Aug 08 |
peter |
36 |
namespace theplu { |
1445 |
27 Aug 08 |
peter |
37 |
namespace yat { |
1497 |
12 Sep 08 |
peter |
38 |
namespace normalizer { |
1445 |
27 Aug 08 |
peter |
39 |
|
1445 |
27 Aug 08 |
peter |
40 |
/** |
1521 |
21 Sep 08 |
peter |
\brief Zero mean and unity variance |
1445 |
27 Aug 08 |
peter |
42 |
|
1717 |
13 Jan 09 |
peter |
Shift and scale the values in a range as: \f$ y_i = |
4048 |
03 Mar 21 |
peter |
\frac{x_i-m}{k*s} \f$ where \a m is the mean, \a s is the |
4048 |
03 Mar 21 |
peter |
standard deviation, and \a k is set in constructor (default: |
4048 |
03 Mar 21 |
peter |
1.0). After normalization, the range will have zero mean and |
4048 |
03 Mar 21 |
peter |
variance = \f$ 1/k^2 \f$. |
1445 |
27 Aug 08 |
peter |
48 |
|
3541 |
23 Dec 16 |
peter |
Type Requirements: |
3541 |
23 Dec 16 |
peter |
- \c InputIterator is \ref concept_data_iterator |
3541 |
23 Dec 16 |
peter |
- \c InputIterator is \forward_traversal_iterator |
3541 |
23 Dec 16 |
peter |
- \c OutputIterator is \writable_iterator |
3541 |
23 Dec 16 |
peter |
- \c OutputIterator is \forward_traversal_iterator |
3541 |
23 Dec 16 |
peter |
54 |
|
1445 |
27 Aug 08 |
peter |
\since New in yat 0.5 |
1445 |
27 Aug 08 |
peter |
56 |
*/ |
1521 |
21 Sep 08 |
peter |
57 |
class Zscore |
1445 |
27 Aug 08 |
peter |
58 |
{ |
1445 |
27 Aug 08 |
peter |
59 |
public: |
1445 |
27 Aug 08 |
peter |
60 |
/** |
4048 |
03 Mar 21 |
peter |
\param k In the scaling step elements are divided by \c k times |
4048 |
03 Mar 21 |
peter |
the standard deviation |
4048 |
03 Mar 21 |
peter |
63 |
|
4048 |
03 Mar 21 |
peter |
\since New in yat 0.19 |
4048 |
03 Mar 21 |
peter |
65 |
*/ |
4049 |
03 Mar 21 |
peter |
66 |
explicit Zscore(double k=1.0); |
4048 |
03 Mar 21 |
peter |
67 |
|
4048 |
03 Mar 21 |
peter |
68 |
/** |
4048 |
03 Mar 21 |
peter |
\return the k-factor set in constructor |
4048 |
03 Mar 21 |
peter |
70 |
|
4048 |
03 Mar 21 |
peter |
\since new in yat 0.19 |
4048 |
03 Mar 21 |
peter |
72 |
*/ |
4048 |
03 Mar 21 |
peter |
73 |
double k(void) const; |
4048 |
03 Mar 21 |
peter |
74 |
|
4048 |
03 Mar 21 |
peter |
75 |
/** |
1521 |
21 Sep 08 |
peter |
The element in range [result, result + (last-first)) is |
1717 |
13 Jan 09 |
peter |
calculated as result[i] = (first[i] - m) / s where m and std |
1521 |
21 Sep 08 |
peter |
are the mean and standard deviation, respectively, of the range |
1521 |
21 Sep 08 |
peter |
[first, last). |
1445 |
27 Aug 08 |
peter |
80 |
|
1445 |
27 Aug 08 |
peter |
It is possible to centralize a range "in place"; it is |
1445 |
27 Aug 08 |
peter |
permissible for the iterators \a first and \a result to be the |
1445 |
27 Aug 08 |
peter |
same. \see std::transform |
1445 |
27 Aug 08 |
peter |
84 |
*/ |
3541 |
23 Dec 16 |
peter |
85 |
template<class InputIterator, class OutputIterator> |
3541 |
23 Dec 16 |
peter |
86 |
void operator()(InputIterator first, InputIterator last, |
3541 |
23 Dec 16 |
peter |
87 |
OutputIterator result) const |
1445 |
27 Aug 08 |
peter |
88 |
{ |
3541 |
23 Dec 16 |
peter |
89 |
BOOST_CONCEPT_ASSERT((utility::DataIteratorConcept<InputIterator>)); |
3541 |
23 Dec 16 |
peter |
90 |
BOOST_CONCEPT_ASSERT((utility::DataIteratorConcept<OutputIterator>)); |
3541 |
23 Dec 16 |
peter |
91 |
|
3541 |
23 Dec 16 |
peter |
// needed for weighted_if_any2 |
3541 |
23 Dec 16 |
peter |
93 |
typename utility::weighted_if_any2<InputIterator, OutputIterator>::type tag; |
1739 |
21 Jan 09 |
peter |
94 |
normalize(first, last, result, tag); |
1524 |
23 Sep 08 |
peter |
95 |
} |
1524 |
23 Sep 08 |
peter |
96 |
|
1524 |
23 Sep 08 |
peter |
97 |
private: |
4048 |
03 Mar 21 |
peter |
98 |
double k_; |
4048 |
03 Mar 21 |
peter |
99 |
|
4048 |
03 Mar 21 |
peter |
100 |
|
2157 |
18 Jan 10 |
peter |
101 |
template<class ForwardIterator, class OutputIterator> |
2157 |
18 Jan 10 |
peter |
102 |
void normalize(ForwardIterator first, ForwardIterator last, |
2157 |
18 Jan 10 |
peter |
103 |
OutputIterator result, |
1739 |
21 Jan 09 |
peter |
104 |
utility::unweighted_iterator_tag tag) const |
1524 |
23 Sep 08 |
peter |
105 |
{ |
3541 |
23 Dec 16 |
peter |
// we need to traverse the input range once in calculating m |
3541 |
23 Dec 16 |
peter |
// (and std) and a second time when assigning new values, so |
3541 |
23 Dec 16 |
peter |
// single pass iterator will not suffice. |
3541 |
23 Dec 16 |
peter |
109 |
BOOST_CONCEPT_ASSERT((boost_concepts::ForwardTraversal<ForwardIterator>)); |
3541 |
23 Dec 16 |
peter |
110 |
|
3541 |
23 Dec 16 |
peter |
111 |
BOOST_CONCEPT_ASSERT((boost_concepts::WritableIterator<OutputIterator>)); |
3541 |
23 Dec 16 |
peter |
112 |
|
1521 |
21 Sep 08 |
peter |
113 |
statistics::Averager a; |
1521 |
21 Sep 08 |
peter |
114 |
add(a, first, last); |
1521 |
21 Sep 08 |
peter |
115 |
double m = a.mean(); |
4048 |
03 Mar 21 |
peter |
116 |
double factor = 1.0 / (a.std() * k_); |
1521 |
21 Sep 08 |
peter |
117 |
while (first!=last) { |
4048 |
03 Mar 21 |
peter |
118 |
*result = (*first - m) * factor; |
1521 |
21 Sep 08 |
peter |
119 |
++first; |
1521 |
21 Sep 08 |
peter |
120 |
++result; |
1521 |
21 Sep 08 |
peter |
121 |
} |
1445 |
27 Aug 08 |
peter |
122 |
} |
1445 |
27 Aug 08 |
peter |
123 |
|
3541 |
23 Dec 16 |
peter |
124 |
template<class ForwardIterator, class OutputIterator> |
3541 |
23 Dec 16 |
peter |
125 |
void normalize(ForwardIterator first, ForwardIterator last, |
3541 |
23 Dec 16 |
peter |
126 |
OutputIterator result, |
1739 |
21 Jan 09 |
peter |
127 |
utility::weighted_iterator_tag tag) const |
1524 |
23 Sep 08 |
peter |
128 |
{ |
3541 |
23 Dec 16 |
peter |
// we need to traverse the input range once in calculating m |
3541 |
23 Dec 16 |
peter |
// (and std) and a second time when assigning new values, so |
3541 |
23 Dec 16 |
peter |
// single pass iterator will not suffice. |
3541 |
23 Dec 16 |
peter |
132 |
BOOST_CONCEPT_ASSERT((boost_concepts::ForwardTraversal<ForwardIterator>)); |
3541 |
23 Dec 16 |
peter |
133 |
|
3541 |
23 Dec 16 |
peter |
134 |
BOOST_CONCEPT_ASSERT((boost_concepts::WritableIterator<OutputIterator>)); |
3541 |
23 Dec 16 |
peter |
// we traverse output range twice |
3541 |
23 Dec 16 |
peter |
136 |
BOOST_CONCEPT_ASSERT((boost_concepts::ForwardTraversal<OutputIterator>)); |
2157 |
18 Jan 10 |
peter |
137 |
detail::copy_weight_if_weighted(first, last, result); |
1524 |
23 Sep 08 |
peter |
138 |
statistics::AveragerWeighted a; |
1524 |
23 Sep 08 |
peter |
139 |
add(a, first, last); |
1524 |
23 Sep 08 |
peter |
140 |
double m = a.mean(); |
4048 |
03 Mar 21 |
peter |
141 |
double factor = 1.0 / (a.std() * k_); |
3541 |
23 Dec 16 |
peter |
142 |
utility::iterator_traits<ForwardIterator> in_trait; |
3541 |
23 Dec 16 |
peter |
143 |
utility::iterator_traits<OutputIterator> out_trait; |
1524 |
23 Sep 08 |
peter |
144 |
while (first!=last) { |
4048 |
03 Mar 21 |
peter |
145 |
out_trait.data(result) = (in_trait.data(first) - m) * factor; |
1524 |
23 Sep 08 |
peter |
146 |
++first; |
1524 |
23 Sep 08 |
peter |
147 |
++result; |
1524 |
23 Sep 08 |
peter |
148 |
} |
1524 |
23 Sep 08 |
peter |
149 |
} |
1445 |
27 Aug 08 |
peter |
150 |
}; |
1445 |
27 Aug 08 |
peter |
151 |
|
1497 |
12 Sep 08 |
peter |
152 |
}}} // end of namespace normalizer, yat and thep |
1445 |
27 Aug 08 |
peter |
153 |
#endif |