Part 5. Modeling dependencies with Bayesian and Markov networks