Page 1 of 1

Spatio-temporal models in runmlwin

Posted: Tue Apr 24, 2018 7:50 pm
by yransome
Has anyone estimated a spatio-temporal model using the runmlwin platform?
I have 6 years of data across areas, and i am wondering how/where to specify the space * time correlation? And how would one print/predict the relative risks for the areas over time to map?
If anyone has syntax and willing to share will also be helpful.
Many thanks in advance!

level3 is where the constant and car weights are specified
level 2 is the area level random effect
level 1 is the observation level random effect for the poisson ofset.

runmlwin alcohol cons year , ///
level3(area: cons, carids(mmid1-mmid10) carweights(carweight1-carweight10)) ///
level2(area: cons ) ///
level1(area: ) ///
discrete(distribution(poisson) link(log) offset(logexp)) ///
mcmc(chain(50000) refresh(500)) initsprevious nopause