Bayesian Networks, 2nd Edition

5

More Complex Cases: General Bayesian Networks

In this chapter we will conclude our exploration of BNs, moving to the more general case in which each variable in the data is modelled with the random variable that best suits it rather than limiting ourselves to multinomial and normal distributions. For this purpose, we will use the Stan (Carpenter et al., 2017) MCMC sampler through its interface rstan (Stan Development Team, 2020b).

5.1 Introductory Example: A & E Waiting Times

Suppose that we are interested in estimating the waiting times in the Accidents & Emergency (A & E) department of a hospital. Much information is publicly available on the subject, since this is one of the key metrics A & E departments are evaluated on. For instance, ...