2964938635

Nodes posses the ability to transmit the information (messages) to all other nodes. Therefore eveiy node will contain necessary information to -4 >estimate the joint probability distribution trough the combining of the eonditional distributions.

•    t

•    •

1.........................................................................................................f..........................................................................................................i.....

◄^Therefore all the methods of improving of the inference effectiveness; seem to be crucial. The increasing of the algorithm effectiveness may bej achieved using the appropriate dispersing of the calculation and storing the particular temporary results (at the cost of increasing of the need of memoiy resources). It can be achieved owing to the distribution of the! ^-►factors belonging to the formula (i) in different branches of the special propagation tree [Michalewicz et al 1993], [K³opotek 1994], [Wierzchonj 1993]- Intuitively, it can be explained as the storing of particular sets of yariables in particular tree nodes. Nodes posses the ability to transmitj the information (messages) to all other nodes. Therefore every node wilH ► contain necessaiy information to estimate the joint probability distribution trough the combining of the eonditional distributions.j Shenoy and Shafer have proposed [Shenoy&Shafer 1990] the methodj owing to which nonę of nodes can count the joint probability distribution on the basis of eonditional distributions and instead uses the boundary distributions of the variables at node.

.....J-.........................................................................................................'.............................................................................................................^.....

Nodes posses the ability to transmit the information (messages) to all other nodes. Therefore every node will contain necessary information to estimate the joint probability distribution trough the combining of the eonditional distributions.