Daniel@0: % Make an HMM with autoregressive Gaussian observations (switching AR model)
Daniel@0: %   X1 -> X2
Daniel@0: %   |     | 
Daniel@0: %   v     v
Daniel@0: %   Y1 -> Y2 
Daniel@0: 
Daniel@0: seed = 0;
Daniel@0: rand('state', seed);
Daniel@0: randn('state', seed);
Daniel@0: 
Daniel@0: intra = zeros(2);
Daniel@0: intra(1,2) = 1;
Daniel@0: inter = zeros(2);
Daniel@0: inter(1,1) = 1;
Daniel@0: inter(2,2) = 1;
Daniel@0: n = 2;
Daniel@0: 
Daniel@0: Q = 2; % num hidden states
Daniel@0: O = 2; % size of observed vector
Daniel@0: 
Daniel@0: ns = [Q O];
Daniel@0: dnodes = 1;
Daniel@0: onodes = [2];
Daniel@0: bnet = mk_dbn(intra, inter, ns, 'discrete', dnodes, 'observed', onodes);
Daniel@0: 
Daniel@0: bnet.CPD{1} = tabular_CPD(bnet, 1);
Daniel@0: bnet.CPD{2} = gaussian_CPD(bnet, 2);
Daniel@0: bnet.CPD{3} = tabular_CPD(bnet, 3);
Daniel@0: bnet.CPD{4} = gaussian_CPD(bnet, 4);
Daniel@0: 
Daniel@0: 
Daniel@0: T = 10; % fixed length sequences
Daniel@0: 
Daniel@0: engine = {};
Daniel@0: %engine{end+1} = hmm_inf_engine(bnet);
Daniel@0: engine{end+1} = jtree_unrolled_dbn_inf_engine(bnet, T);
Daniel@0: %engine{end+1} = smoother_engine(hmm_2TBN_inf_engine(bnet));
Daniel@0: %engine{end+1} = smoother_engine(jtree_2TBN_inf_engine(bnet));
Daniel@0: 
Daniel@0: inf_time = cmp_inference_dbn(bnet, engine, T, 'check_ll',1);
Daniel@0: learning_time = cmp_learning_dbn(bnet, engine, T, 'check_ll', 1);
Daniel@0: