annotate toolboxes/FullBNT-1.0.7/netlab3.3/demgpot.m @ 0:cc4b1211e677 tip

initial commit to HG from Changeset: 646 (e263d8a21543) added further path and more save "camirversion.m"
author Daniel Wolff
date Fri, 19 Aug 2016 13:07:06 +0200
parents
children
rev   line source
Daniel@0 1 function g = demgpot(x, mix)
Daniel@0 2 %DEMGPOT Computes the gradient of the negative log likelihood for a mixture model.
Daniel@0 3 %
Daniel@0 4 % Description
Daniel@0 5 % This function computes the gradient of the negative log of the
Daniel@0 6 % unconditional data density P(X) with respect to the coefficients of
Daniel@0 7 % the data vector X for a Gaussian mixture model. The data structure
Daniel@0 8 % MIX defines the mixture model, while the matrix X contains the data
Daniel@0 9 % vector as a row vector. Note the unusual order of the arguments: this
Daniel@0 10 % is so that the function can be used in DEMHMC1 directly for sampling
Daniel@0 11 % from the distribution P(X).
Daniel@0 12 %
Daniel@0 13 % See also
Daniel@0 14 % DEMHMC1, DEMMET1, DEMPOT
Daniel@0 15 %
Daniel@0 16
Daniel@0 17 % Copyright (c) Ian T Nabney (1996-2001)
Daniel@0 18
Daniel@0 19 % Computes the potential gradient
Daniel@0 20
Daniel@0 21 temp = (ones(mix.ncentres,1)*x)-mix.centres;
Daniel@0 22 temp = temp.*(gmmactiv(mix,x)'*ones(1, mix.nin));
Daniel@0 23 % Assume spherical covariance structure
Daniel@0 24 if ~strcmp(mix.covar_type, 'spherical')
Daniel@0 25 error('Spherical covariance only.')
Daniel@0 26 end
Daniel@0 27 temp = temp./(mix.covars'*ones(1, mix.nin));
Daniel@0 28 temp = temp.*(mix.priors'*ones(1, mix.nin));
Daniel@0 29 g = sum(temp, 1)/gmmprob(mix, x);