Point predictions vs posterior predictive checking

This week I had a cool question come across my desk at work: if the density of model point predictions doesn’t look like the density of outcomes, is something wrong? My mate Dave also asked me about posterior predictive checking. So I wrote a thing about both.

Posted on Rpubs because it deals with code so much more easily.

http://rpubs.com/jimsavage/posterior_predictive_checking

Comments are closed.