1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
|
function retval = corr(x, y) % --*-- Unitary tests --*--
%@info:
%! @deftypefn {Function File} {} corr (@var{x})
%! @deftypefnx {Function File} {} corr (@var{x}, @var{y})
%! Compute matrix of correlation coefficients.
%! @anchor{corr}
%!
%! If each row of @var{x} and @var{y} is an observation and each column is
%! a variable, then the @w{(@var{i}, @var{j})-th} entry of
%! @code{corr (@var{x}, @var{y})} is the correlation between the
%! @var{i}-th variable in @var{x} and the @var{j}-th variable in @var{y}.
%! @tex
%! $$
%! {\rm corr}(x,y) = {{\rm cov}(x,y) \over {\rm std}(x) {\rm std}(y)}
%! $$
%! @end tex
%! @ifnottex
%!
%! @example
%! corr (x,y) = cov (x,y) / (std (x) * std (y))
%! @end example
%!
%! @end ifnottex
%! If called with one argument, compute @code{corr (@var{x}, @var{x})},
%! the correlation between the columns of @var{x}.
%! @end deftypefn
%@eod:
%
% Notes: - the original Octave code has been rewritten to avoid calling cov, since
% there is a long-standing incompatiblity between Matlab's cov and Octave's cov
% (see https://savannah.gnu.org/bugs/?40751)
% - For compatibility with Matlab, the correlation of a constant
% is defined as NaN, not 1
%
% Adapted for Matlab (R) from GNU Octave 4.0.1
% Original files: statistics\base\corr.m, statistics\base\cov.m, and packages\stk-2.3.4\misc\mole\corr\corr.m
% Original authors: Kurt Hornik <hornik@wu-wien.ac.at> and Julien Bect <julien.bect@supelec.fr>
% Copyright (C) 1993-1996 Kurt Hornik
% Copyright (C) 1996-2015 John W. Eaton
% Copyright (C) 2013-2015 Julien Bect
% Copyright (C) 2016-2017 Dynare Team
%
% This file is part of Dynare.
%
% Dynare is free software: you can redistribute it and/or modify
% it under the terms of the GNU General Public License as published by
% the Free Software Foundation, either version 3 of the License, or
% (at your option) any later version.
%
% Dynare is distributed in the hope that it will be useful,
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
% GNU General Public License for more details.
%
% You should have received a copy of the GNU General Public License
% along with Dynare. If not, see <http://www.gnu.org/licenses/>.
if (nargin < 1 || nargin > 2)
error('corr needs to be called with 1 or 2 input arguments')
end
% Input validation
if (nargin==1 && ~(isnumeric (x) || islogical (x))) || ...
(nargin==2 && ~(isnumeric (x) || islogical (x) || isnumeric (y) || islogical (y)))
error ('corr: X and Y must be numeric matrices or vectors');
end
if (nargin==1 && ~ismatrix(x)) || (nargin==2 && (~ismatrix(y) || ~ismatrix(x)))
error ('corr: X and Y must be 2-D matrices or vectors');
end
if (nargin == 2)
if size(y,1)~=size(x,1)
error('corr: X and Y must have the same length')
end
end
% Special case, correlation with scalar is NaN in Matlab
if isscalar(x)
if nargin==1
retval = NaN;
else
retval = NaN(size(y));
end
return
end
if nargin==2 && isscalar(y)
retval = NaN(size(x'));
return
end
n = size (x, 1);
x = x - repmat(mean(x),n,1); %demean
sx = std(x,1); %standard deviation
sx(sx==0)=NaN; %take care of constant vectors
if (nargin == 2)
y = y - repmat (mean (y), n, 1);
sy = std (y, 1);
sy (sy == 0) = nan;
else
y = x;
sy = sx;
end
c = x'*y;
s = sx'*sy;
retval = c./(n * s);
end
%@test:1
%$ x = rand (10);
%$ cc1 = corr(x);
%$ cc2 = corr(x, x);
%$ t(1) = dassert(size(cc1),[10, 10]);
%$ t(2) = dassert(size (cc2),[10, 10]);
%$ t(3) = dassert(cc1, cc2, sqrt (eps));
%$ T = all(t);
%@eof:1
%@test:2
%$ x = [1:3]';
%$ y = [3:-1:1]';
%$ t = zeros(3,1);
%$ t(1) = dassert(corr(x, y), -1, 5*eps);
%$ t(2) = dassert(corr(x, flipud (y)), 1, 5*eps);
%$ t(3) = dassert(corr([x, y]), [1 -1; -1 1], 5*eps);
%$ T = all(t);
%@eof:2
%@test:3
%$ if ~isoctave()
%$ t = zeros(3,1);
%$ t(1) = dassert(corr(5), NaN);
%$ t(2) = dassert(corr([1 2 3],5),NaN(3,1));
%$ t(3) = dassert(corr(5,[1 2 3]),NaN(1,3));
%$ else
%$ t = 1;
%$ end
%$ T = all(t);
%@eof:3
%@test:4
%$ t = zeros(6,1);
%$ try
%$ corr()
%$ t(1) = false;
%$ catch
%$ t(1) = true;
%$ end
%$ try
%$ corr(1, 2, 3)
%$ t(2) = false;
%$ catch
%$ t(2) = true;
%$ end
%$ try
%$ corr([1; 2], ['A', 'B'])
%$ t(3) = false;
%$ catch
%$ t(3) = true;
%$ end
%$ try
%$ corr([1; 2], ['A', 'B'])
%$ t(4) = false;
%$ catch
%$ t(4) = true;
%$ end
%$ try
%$ corr(ones (2,2,2))
%$ t(5) = false;
%$ catch
%$ t(5) = true;
%$ end
%$ try
%$ corr(ones (2,2), ones (2,2,2))
%$ t(6) = false;
%$ catch
%$ t(6) = true;
%$ end
%$ T = all(t);
%@eof:4
|
