Learning dynamic Bayesian networks with mixed variables

This project considers dynamic Bayesian networks for discrete and continuous variables. We restrict us to the case, where the distributionof the variables is conditional Gaussian. We show how to learn the parameters in and the structure of a dynamic Bayesian network and also how the Markov order can be learned. An automated procedure for specifying prior distributions for the parameters in a dynamic Bayesian network is presented. It is a simple extension of the procedure for the ordinary Bayesian networks. Finally the Wölfer's sunspot numbers are analysed.