Learning with Bayesian netowrk

In this project we have built and learned different Bayesian networks for car crash dataset. In general we have constructed 7 different networks with different modifications and improvements.

Things are covered:

• Data preparation

• BN structure learning with the use of chi-squared test (manual)

• BN structure learning with complete and incomplete observations

• BN improvement with MCMC and K2

• BN structure learning with MCMC and EM

• Inference ilustration

• In total 7 BNs with corresponding Log likelihood, BIC and Bayesian scores.