Dear,
I have idea to solve cyclycal bayesian networks by transforming
them into a syngly.But question is what to do with distribution...??
And I think that when you transform cyclycal B. net you must
rebuild distributions manualy(that's becouse you are making an
adequate model).
Nodes must be transformed using Demster-Sheifer methods,
so if you have a symple net vith two parents(A,B) and child C,
you must add nodes(A or B) and (A and B) and rebuild
distributions for A, B, (A and B), (A or B)...!!
And every multiple connected networks ends in NOISY OR...!!
Montecarlo symulation algorythms doesent give a correct solution,
they only find solution which is a balance betwen propagation
(causal, and diagnostyc), but not real solution..
I would try to implement B.Nets in my firm, so if you re interesed,
in solving this in LISP(My favorite language) you can step into a conntact..!!
Thanks in advance, Robert ..!!
tel:38522312027
To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre.
_______________________________________________ uai mailing list [email protected] https://secure.engr.oregonstate.edu/mailman/listinfo/uai
