Bayesian Network Fundamentals
A Bayesian network is:
- A directed acyclic graph (DAG)
G whose nodes represent the random variablesX1,X2,...,Xn - For each node
Xi a CPDP(Xi|ParG(Xi))
- A directed acyclic graph (DAG)
The BN represents a joint distribution via the chain rule for Bayesian Networks
P(X1,X2,...,Xn)=∏iP(Xi|ParG(Xi)) BN is a legal distribution:
P(X1,X2,...,Xn)≥0
∑iP(X1,X2,...,Xn)=1 P factorized overG - Let
G be a graph overX1,X2,...,Xn P factorized overG if
P(X1,X2,...,Xn)=∏iP(Xi|ParG(Xi))
- Let
Reasoning patterns:
- Causal reasoning
- Evidential reasoning
- Intercausal reasoning
Active Trail
- A trail
X1−X2−...−Xk is active if it has no v-structuresXi−1→Xi←Xi+1 - A trail
X1−X2−...−Xk is active givenZ if:
- for any v-structure
Xi−1→Xi←Xi+1 , we have thatXi or one of its descendants∈Z - no other
Xi is inZ
- for any v-structure
- A trail