Assigning max likelihood to one state and Back propagation of BN

[JOSELUIS]

Hello

I have a problem in my VB Macro (linking a BN (domain.NET) and an Excell sheet).

When I try to impose a likelihood, in this case 100% for the upper state of a child node, I have two problems:
- The new saved domain does not incorporate that likelihood

- When I include a new parent node linked with that child node, to make it maximum for the upper state, the back propagation is not well done in my new saved domain (.NET).
 
I don´t know the reason for that.
My domain just include discrete chance nodes

Does anybody know the code for that o how to do it?

Thanks in advance

[Anders L Madsen]

Hi,

I'm not sure that I understand what younare aiming at doing.

The .net files do not include evidence. It only contains a specification of the network. You can store evidence sets in different "cases" files (saveCase).

Belief update and adaptation of parameters are described in detail in the HUGIN API Reference Manual (using the C API as example)