1
$\begingroup$

This is a simple question on Bayesian spatial modelling via conditional autoregressive modelling.

What is, according to your judgement (and possibly some methodological source), the minimum number of regions for which CAR is suitable? Would you say that with a low number of regions, say 9, it is highly advisable to use a model with as little parameters as possible, that is, the intrinsic model instead of the better performing Leroux specification?

Reference: Lee 2011: A Comparison of Conditional Autoregressive Models Used in Bayesian Disease Mapping, Spatial and Spatio-temporal Epidemiology 2/2, p 79–89

Improve this question
$\endgroup$

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.