wolffd@0: 
wolffd@0: <h2>Example of more complex parameter tieing</h2>
wolffd@0: 
wolffd@0: We now give a more complex pattern of parameter tieing.
wolffd@0: (This example is due to Rainer Deventer.)
wolffd@0: The structure is as follows:
wolffd@0: <p>
wolffd@0: <img src="Figures/rainer_tied.gif" height="400">
wolffd@0: <!--<img src="rainer_dbn.jpg" height="600">-->
wolffd@0: <p>
wolffd@0: Since nodes 2 and 3 in slice 2 (N7 and N8)
wolffd@0: have different parents than their counterparts in slice 1 (N2 and N3),
wolffd@0: they must be put into different equivalence classes.
wolffd@0: Hence we define
wolffd@0: <pre>
wolffd@0: eclass1 = [1 2 3 4 5];
wolffd@0: eclass2 = [1 6 7 4 5];
wolffd@0: </pre>
wolffd@0: The dotted bubbles represent the equivalence classes.
wolffd@0: Node 7 is the representative node for equivalence class
wolffd@0: 6, and node 8 is the rep. for class 7, so we need to write
wolffd@0: <pre>
wolffd@0: bnet.CPD{6} = xxx_CPD(bnet, 7, xxx);
wolffd@0: bnet.CPD{7} = xxx_CPD(bnet, 8, xxx);
wolffd@0: </pre>
wolffd@0: In general, you can use the following code fragment:
wolffd@0: <pre>
wolffd@0: eclass = bnet.equiv_class(:);
wolffd@0: for e=1:max(eclass)
wolffd@0:   i = bnet.rep_of_eclass(e);
wolffd@0:   bnet.CPD{e} = xxx_CPD(bnet,i);
wolffd@0: end
wolffd@0: </pre>
wolffd@0: <!--
wolffd@0: which is equivalent to
wolffd@0: <pre>
wolffd@0: E = max(eclass);
wolffd@0: rep = zeros(1,E);
wolffd@0: for e=1:E
wolffd@0:   mems = find(eclass==e);
wolffd@0:   rep(e) = mems(1);
wolffd@0: end
wolffd@0: for e=1:E
wolffd@0:   bnet.CPD{e} = xxx_CPD(bnet, rep(e));
wolffd@0: end
wolffd@0: </pre>
wolffd@0: -->