wolffd@0: function bnet = learn_params_dbn(bnet, data)
wolffd@0: % LEARN_PARAM_DBN Estimate params of a DBN for a fully observed model
wolffd@0: % bnet = learn_params_dbn(bnet, data)
wolffd@0: %
wolffd@0: % data(i,t) is the value of node i in slice t (can be a cell array)
wolffd@0: % We currently assume there is a single time series
wolffd@0: %
wolffd@0: % We set bnet.CPD{i} to its ML/MAP estimate.
wolffd@0: %
wolffd@0: % Currently we assume each node in the first 2 slices has its own CPD (no param tying);
wolffd@0: % all nodes in slices >2 share their params with slice 2 as usual.
wolffd@0: 
wolffd@0: [ss T] = size(data);
wolffd@0: 
wolffd@0: % slice 1
wolffd@0: for j=1:ss
wolffd@0:   if adjustable_CPD(bnet.CPD{j})
wolffd@0:     fam = family(bnet.dag,j);
wolffd@0:     bnet.CPD{j} = learn_params(bnet.CPD{j}, data(fam,1));
wolffd@0:   end
wolffd@0: end
wolffd@0: 
wolffd@0: 
wolffd@0: % slices 2:T
wolffd@0: % data2(:,t) contains [data(:,t-1); data(:,t)].
wolffd@0: % Then we extract out the rows corresponding to the parents in the current and previous slice.
wolffd@0: data2 = [data(:,1:T-1);
wolffd@0: 	 data(:,2:T)];
wolffd@0: for j=1:ss
wolffd@0:   j2 = j+ss;
wolffd@0:   if adjustable_CPD(bnet.CPD{j2})
wolffd@0:     fam = family(bnet.dag,j2);
wolffd@0:     bnet.CPD{j2} = learn_params(bnet.CPD{j2}, data2(fam,:));
wolffd@0:   end
wolffd@0: end
wolffd@0: