*Bayesian Networks (BN)
Defi: A BN is a triplet (N,D,P) where
- N is a set of random variables,
- D is a directed acyclic graph (DAG) whose nodes are labeled by elements of N,
- P is a jpd over N specified by probability distributions of each node x in D conditioned on its parents parn(x) in D.
D expresses dependency relations among elements of N.
- A variable is independent of its non-descendents given its parents.
- Hence P can be expressed as P(N) = ? x? N P(x | parn(x)).