Mercurial > hg > camir-aes2014

annotate toolboxes/FullBNT-1.0.7/KPMstats/dirichletrnd.m @ 0:e9a9cd732c1e tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
rev   line source
wolffd@0 1 function x = dirichletrnd(alpha)
wolffd@0 2 %DIRICHLETRND Random vector from a dirichlet distribution.
wolffd@0 3 % x = dirichletrnd(alpha) returns a vector randomly selected
wolffd@0 4 % from the Dirichlet distribution with parameter vector alpha.
wolffd@0 5 %
wolffd@0 6 % The algorithm used is the following:
wolffd@0 7 % For each alpha(i), generate a value s(i) with distribution
wolffd@0 8 % Gamma(alpha(i),1). Now x(i) = s(i) / sum_j s(j).
wolffd@0 9 %
wolffd@0 10 % The above algorithm was recounted to me by Radford Neal, but
wolffd@0 11 % a reference would be appreciated...
wolffd@0 12 % Do the gamma parameters always have to be 1?
wolffd@0 13 %
wolffd@0 14 % Author: David Ross
wolffd@0 15 % $Id: dirichletrnd.m,v 1.1.1.1 2005/05/22 23:32:12 yozhik Exp $
wolffd@0 16
wolffd@0 17 %-------------------------------------------------
wolffd@0 18 % Check the input
wolffd@0 19 %-------------------------------------------------
wolffd@0 20 error(nargchk(1,1,nargin));
wolffd@0 21
wolffd@0 22 if min(size(alpha)) ~= 1 | length(alpha) < 2
wolffd@0 23 error('alpha must be a vector of length at least 2');
wolffd@0 24 end
wolffd@0 25
wolffd@0 26
wolffd@0 27 %-------------------------------------------------
wolffd@0 28 % Main
wolffd@0 29 %-------------------------------------------------
wolffd@0 30 gamma_vals = gamrnd(alpha, ones(size(alpha)), size(alpha));
wolffd@0 31 denom = sum(gamma_vals);
wolffd@0 32 x = gamma_vals / denom;