Skip to content
Prev 28008 / 29559 Next

credible interval for empirical Bayesian estimates of rates

Thanks Roger and list.

I didn't think a repex was needed because a question was: why does
spdep::EBest(counts, population, family = 'binomial') give the same
results at GeoDa's, while EBest(.. binomial) is "binomial" while GeoDa
calls that "Poisson-Gamma". GeoDa can't give use a repex (GUI) and think
this is a question about terminology. The same results were achieved with
the packages while naming the model differently - why?

Yes ?spdep::EBest directed me to the literature I'm struggling to access.
And Yes, I've been looking at the raw code and understand how the estmm is
generated.

I've been using the epitools::pois.exact() and spdep::EBest. I can compare
the point estimates from pois.exact to those provided by EBest, but I'd
like to graph side by side their credible / confidence intervals.

Its this last part on the credible intervals I'm interested in. How to get
credible intervals around estmm? This is my main question.

ASDAR is a reference I'm using all the time. Thanks for that gem.

DCluster::empbaysmooth also does not provide a credible interval, either.

-Dexter
http://dexterlocke.com/
On Fri, Apr 24, 2020 at 10:23 AM Roger Bivand <Roger.Bivand at nhh.no> wrote: