Is there a way to set a condition on passing a parent's node state to its son?

[PCherubini]

I'm modeling a forensic case with genie (see image for a provisional section of the model; O is a suspect and V is the victim):

In node "blood in O car" value "false" means there's no blood in the suspect's car... and of course if there's no blood the "quality of stain" node cannot be neither "pure" nor "diluted", and I'm forced to add a "absent" state to it, only to assign it probability 1 for any combination of value where "blood in O car" is false (similar problem in the node "DNA test": if no blood, no match can be found... but it means an absence of test, not a test with a mismatch outcome). It's a useless complication, and it makes the net less intuitive and less readable for non-experts (and most judges, attorneys, and lawyers, let alone jurors, already find it very hard to understand DAGs: so making them as simple as possible to follow is important).
Is there some way to assign filtering conditions on an arrow? That is, there is a way to instruct the net to pass only the value "true" of "blood in O car"to its sons, and not the probability mass corresponding to the value "false"?
Thank you

[marek [BayesFusion]]

Hi Paolo,

What you would like to be able to model is asymmetry in the conditional probability distribution, something that is quite naturally modeled in probability trees. Unfortunately, this is not very natural for graphical models, like Bayesian networks. There was some work on this in early 1990s (some names to search for are Smith, Holtzman, Matheson, Fishbeck, if I remember correctly) but this work has not been widely adopted and, to my knowledge, has not made it into any practical software. GeNIe does not have any facilities for conditional relationships. What you did in your model is quite likely the best way of handling this problem. Sorry for not being able to give you a positive answer :-(.

Marek

[PCherubini]

Thank you, nevertheless
p