network structure is too complex

The front end.
Post Reply
bahman
Posts: 21
Joined: Sun Mar 13, 2022 12:24 pm

network structure is too complex

Post by bahman »

HI;

in this model NODE13 has 10 parents and each paren 2 state . so NODE13 has 1024 state. but when i want to update i get this error
Error when running Cooper's algorithm
Error (-42): UpdateBeliefs failed - network structure is too complex.
2022-04-21_1-43-43.jpg
2022-04-21_1-43-43.jpg (163.89 KiB) Viewed 1824 times
i tried all of the algorithms. some of them worls but not accurete at all and not calculate other nodes.

i tried target node too but still not work. is it really complex ?

what should i do ? can you suggest another way for define NODE13 ? i mean add another node between this and parents or such an idea !!!!

i will send my model to you Mr Marek and i would be greatful if you check it.

thank you
marek [BayesFusion]
Site Admin
Posts: 430
Joined: Tue Dec 11, 2007 4:24 pm

Re: network structure is too complex

Post by marek [BayesFusion] »

Hi Bahman,

I have looked at your network and it is not really Node13 that is the culprit. Please note that your network will not update even if you remove this node. Generally, it is not that easy to predict the size of the junction tree in advance without trying triangulation. Your network is quite densely connected and even though GeNIe is very efficient, apparently it finds out that the size of the junction tree exceeds available memory. If you insist on this structure, I suggest EPIS-BN, which is quite likely the best sampling algorithm available. It updated the network without evidence on my laptop in 30 seconds with the default 10K samples.

I have noticed that your network has quite a number of dimmed arcs, which means that there is something wrong with the numerical parameters of the child nodes. I assume it is only temporary at this model building stage? You should tend to these definitions at some point.

I hope this helps,

Marek
Post Reply