Bayesian Methods in Health Economics

120 Bayesian Methods in Health Economics

If the model considers more than just one variable to be initialised, it is

easy to extend the definition of the function inits by including in the list

comma-separated suitable values for the other variables. For example, if the

model included a further unobserved random node mu represented by a vector

of size 10 and associated with a Normal distribution, then a suitable code

would be:

inits <- function(){list=(theta=runif(1),mu=rnorm(10))}

In this case, R would generate 10 values from a standard Normal distribution

thatwouldbeusedbyJAGS to initialise the node mu.

The last part of the pre-processing operation