Easter Trailer: Stan is coming
No Stata code this week because of the Easter break, so I thought that I would take the opportunity to trail a couple of the topics that I will be covering over the coming months. It is unusal for me to plan ahead in this way but I have been investigating a Bayesian analysis program […]
Poisson regression with two random effects: MCMC by data augmentation
When I restarted this blog after the Christmas break I resolved to concentrate more on straightforward applications of Bayesian analysis with Stata; I believe that is what most Stata users are interested in. A couple of months in and I am going to ignore my good intensions and discuss a topic that I considered too […]
Poisson regression with two random effects: Better Mata code
Last time I presented a basic Mata program for fitting a Poisson model with two random effects to some epilepsy data. The program was faster that its Stata equivalent but slower than WinBUGS and the mixing was quite poor. This week I will try to improve the performance of the Mata program. It would make sense […]
Poisson regression with two random effects: Mata
In recent postings I have analysed the epilepsy trial of Thall and Vail using Stata, WinBUGS, OpenBUGS and JAGS and after a brief diversion last week I now want to move on to analysing the data using Mata. Previously we saw that an analysis in Stata takes about 6 minutes, while the BUGS-like programs take […]
Is blogging better than publishing?
This week I started with the intension of writing about the use of Mata for analysing the Poisson regression problem that has occupied this blog for the last month. Before preparing the Mata program, I went back to my Stata code to use it as a guide and when I did so, I noticed that […]
Recent Comments