blob: b426fe3091c41246beeb891d407673cc61874cde (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
|
<h2>Example of more complex parameter tieing</h2>
We now give a more complex pattern of parameter tieing.
(This example is due to Rainer Deventer.)
The structure is as follows:
<p>
<img src="Figures/rainer_tied.gif" height="400">
<!--<img src="rainer_dbn.jpg" height="600">-->
<p>
Since nodes 2 and 3 in slice 2 (N7 and N8)
have different parents than their counterparts in slice 1 (N2 and N3),
they must be put into different equivalence classes.
Hence we define
<pre>
eclass1 = [1 2 3 4 5];
eclass2 = [1 6 7 4 5];
</pre>
The dotted bubbles represent the equivalence classes.
Node 7 is the representative node for equivalence class
6, and node 8 is the rep. for class 7, so we need to write
<pre>
bnet.CPD{6} = xxx_CPD(bnet, 7, xxx);
bnet.CPD{7} = xxx_CPD(bnet, 8, xxx);
</pre>
In general, you can use the following code fragment:
<pre>
eclass = bnet.equiv_class(:);
for e=1:max(eclass)
i = bnet.rep_of_eclass(e);
bnet.CPD{e} = xxx_CPD(bnet,i);
end
</pre>
<!--
which is equivalent to
<pre>
E = max(eclass);
rep = zeros(1,E);
for e=1:E
mems = find(eclass==e);
rep(e) = mems(1);
end
for e=1:E
bnet.CPD{e} = xxx_CPD(bnet, rep(e));
end
</pre>
-->
|
