Slides-1, Slides-2. 14.1. Representing Knowledge in an Uncertain Domain: Introduction to Bayesian Network 14.2. The Semantics of Bayesian Networks Representing the full joint distribution A method for constructing Bayesian networks Compactness and node ordering Conditional independence relations in Bayesian networks: Markov Blanket 14.3. Efficient Representation of Conditional Distributions Bayesian nets with continuous variables Continuous variables with discrete and continuous parents 14.4. Exact Inference in Bayesian NetworksDiscrete variables with continuous parents: probit distribution or logit distribution Inference by enumeration The variable elimination algorithm: avoid repeat computation The complexity of exact inference: singly connected or mulltiply connectede Clustering algorithms 14.5. Approximate Inference in Bayesian Networks Direct sampling methods Rejection sampling in Bayesian networks Likelihood weighting Inference by Markov chain simulation The MCMC algorithm Why MCMC works |