Daniel@0: function L = log_lik_complete(bnet, cases, clamped)
Daniel@0: % LOG_LIK_COMPLETE Compute sum_m sum_i log P(x(i,m)| x(pi_i,m), theta_i) for a completely observed data set
Daniel@0: % L = log_lik_complete(bnet, cases, clamped)
Daniel@0: %
Daniel@0: % If there is a missing data, you must use an inference engine.
Daniel@0: % cases(i,m) is the value assigned to node i in case m.
Daniel@0: % (If there are vector-valued nodes, cases should be a cell array.)
Daniel@0: % clamped(i,m) = 1 if node i was set by intervention in case m (default: clamped = zeros)
Daniel@0: % Clamped nodes contribute a factor of 1.0 to the likelihood.
Daniel@0: 
Daniel@0: if iscell(cases), usecell = 1; else usecell = 0; end
Daniel@0: 
Daniel@0: n = length(bnet.dag);
Daniel@0: ncases = size(cases, 2);
Daniel@0: if n ~= size(cases, 1)
Daniel@0:   error('data should be of size nnodes * ncases');
Daniel@0: end
Daniel@0: 
Daniel@0: if nargin < 3, clamped = zeros(n,ncases); end
Daniel@0: 
Daniel@0: L = 0;
Daniel@0: for i=1:n
Daniel@0:   ps = parents(bnet.dag, i);
Daniel@0:   e = bnet.equiv_class(i);
Daniel@0:   u = find(clamped(i,:)==0);
Daniel@0:   L = L + log_prob_node(bnet.CPD{e}, cases(i,u), cases(ps,u));
Daniel@0: end
Daniel@0: