A Let X Be A Node In A Markov Network H And Let Y Be An Assignment Of Values To X S 2528160

/in /by

a. Let X be a node in a Markov network H, and let y be an assignment of values to X’s Markov blanket Y = MBH(X). Provide an efficient algorithm for computing

for x, x’ ∈ Val(X). Your algorithm should involve only local parameters — potentials for cliques involving X.

b. Show how this algorithm applies to Bayesian networks.