Bayesian Networks, 2nd Edition

3

The Mixed Case: Conditional Gaussian Bayesian Networks

In Chapters 1 and 2 we considered BNs with either discrete or continuous variables. Moreover, in each BN all variables followed probability distributions belonging to the same family: multinomial or normal.

In this chapter, we will cover how these two families can be combined to create a conditional Gaussian BN (CGBN). A CGBN is a “mixture of normals” model in which continuous nodes can have both continuous and discrete parents, while discrete nodes can only have discrete parents. We see this as an initial step towards the more complex BNs presented in Chapters 4 and 5, which provide even greater flexibility.

3.1 Introductory Example: Healthcare Costs

The cost of healthcare is a ...