wolffd@0: % Check sampling on a mixture of experts model
wolffd@0: %
wolffd@0: %  X \
wolffd@0: %  | |
wolffd@0: %  Q |
wolffd@0: %  | /
wolffd@0: %  Y
wolffd@0: %
wolffd@0: % where all arcs point down.
wolffd@0: % We condition everything on X, so X is a root node. Q is a softmax, and Y is a linear Gaussian.
wolffd@0: % Q is hidden, X and Y are observed.
wolffd@0: 
wolffd@0: X = 1;
wolffd@0: Q = 2;
wolffd@0: Y = 3;
wolffd@0: dag = zeros(3,3);
wolffd@0: dag(X,[Q Y]) = 1;
wolffd@0: dag(Q,Y) = 1;
wolffd@0: ns = [1 2 2];
wolffd@0: dnodes = [2];
wolffd@0: bnet = mk_bnet(dag, ns, dnodes);
wolffd@0: 
wolffd@0: x = 0.5;
wolffd@0: bnet.CPD{1} = root_CPD(bnet, 1, x);
wolffd@0: bnet.CPD{2} = softmax_CPD(bnet, 2);
wolffd@0: bnet.CPD{3} = gaussian_CPD(bnet, 3);
wolffd@0: 
wolffd@0: data_case = sample_bnet(bnet, 'evidence', {0.8, [], []})
wolffd@0: ll = log_lik_complete(bnet, data_case)
wolffd@0: 
wolffd@0: data_case = sample_bnet(bnet, 'evidence', {-11, [], []})
wolffd@0: ll = log_lik_complete(bnet, data_case)
wolffd@0: 
wolffd@0: