wolffd@0: if 1
wolffd@0:   O = 4;
wolffd@0:   T = 10;
wolffd@0:   nex = 50;
wolffd@0:   M = 2;
wolffd@0:   Q = 3;
wolffd@0: else
wolffd@0:   O = 8;          %Number of coefficients in a vector 
wolffd@0:   T = 420;         %Number of vectors in a sequence 
wolffd@0:   nex = 1;        %Number of sequences 
wolffd@0:   M = 1;          %Number of mixtures 
wolffd@0:   Q = 6;          %Number of states 
wolffd@0: end
wolffd@0: cov_type = 'full';
wolffd@0: 
wolffd@0: data = randn(O,T,nex);
wolffd@0: 
wolffd@0: % initial guess of parameters
wolffd@0: prior0 = normalise(rand(Q,1));
wolffd@0: transmat0 = mk_stochastic(rand(Q,Q));
wolffd@0: 
wolffd@0: if 0
wolffd@0:   Sigma0 = repmat(eye(O), [1 1 Q M]);
wolffd@0:   % Initialize each mean to a random data point
wolffd@0:   indices = randperm(T*nex);
wolffd@0:   mu0 = reshape(data(:,indices(1:(Q*M))), [O Q M]);
wolffd@0:   mixmat0 = mk_stochastic(rand(Q,M));
wolffd@0: else
wolffd@0:   [mu0, Sigma0] = mixgauss_init(Q*M, data, cov_type);
wolffd@0:   mu0 = reshape(mu0, [O Q M]);
wolffd@0:   Sigma0 = reshape(Sigma0, [O O Q M]);
wolffd@0:   mixmat0 = mk_stochastic(rand(Q,M));
wolffd@0: end
wolffd@0: 
wolffd@0: [LL, prior1, transmat1, mu1, Sigma1, mixmat1] = ...
wolffd@0:     mhmm_em(data, prior0, transmat0, mu0, Sigma0, mixmat0, 'max_iter', 5);
wolffd@0: 
wolffd@0: 
wolffd@0: loglik = mhmm_logprob(data, prior1, transmat1, mu1, Sigma1, mixmat1);
wolffd@0: