File: MultiNormalDraw.m

package info (click to toggle)
psychtoolbox-3 3.0.19.14.dfsg1-1
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 86,796 kB
  • sloc: ansic: 176,245; cpp: 20,103; objc: 5,393; sh: 2,753; python: 1,397; php: 384; makefile: 193; java: 113
file content (32 lines) | stat: -rw-r--r-- 798 bytes parent folder | download | duplicates (7)
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
function y = MultiNormalDraw(n,u,K)
% x = MultiNormalDraw(n,u,K)
%
% Make n multivariate normal draws with mean u and covariance matrix K.
% Each draw is in a single column of y, which has n columns.
%
% The routine operates by computing the appropriate linear transformation
% of a N(0,I) multivariate normal draw.  This transformation is given by 
% y= C'x + u where K = C'C.  This works because the covariance of a
% distribution y = Cx is in general given by Ky = C Kx C'.  In our case 
% Kx= I so Ky = C'C = K.
%
% K = 0 is handled as a special case

% 12/31/93  dhb     Added this line.
% 7/24/04   awi     Cosmetic.


[m,null] = size(u);
x = zeros(m,n);
if (MatMax(K) == 0)
  C = zeros(m,m);
else
  C = chol(K);
end

for i = 1:n
  x(:,i) = NormalDraw(m,0,1);
  y(:,i) = C'*x(:,i) + u;
end