Re: [R] Bayesian logistic regression with a beta prior (MCMClogit)

From: francogrex <francogrex_at_mail.com>
Date: Thu, 03 May 2007 13:49:05 -0700 (PDT)

Hi, yes but I realized afterwards that it's the logfun argument that had to be put to logfun=F and the logpriorfun function had to be log=F logpriorfun <- function(beta,shape1,shape2){ sum(dbeta(beta,shape1,shape2,log=F)) }

But that's just for that particular example. I find I am having problems still even after adjusting for that. Using other data it is not accepting the estimation of beta.start with maximum likelihood ("user.prior.density(beta.start) <= 0") and it is obliging me to specify it giving me a very narrow range, and hence the acceptance rate of the output is very mediocre (0.01)... I don't know I am missing something here maybe. As much as I was excited about the MCMCpack, I am finding that it is no substitute for BUGS/Brugs.

Cody_Hamilton wrote:
>
> Dear Franco,
>
> Have you tried using the beta.start option in MCMClogit? (The problem may
> be where you are starting your chain.)
>
> Regards,
> -Cody
>

-- 
View this message in context: http://www.nabble.com/Bayesian-logistic-regression-with-a-beta-prior-%28MCMClogit%29-tf3684970.html#a10312824
Sent from the R help mailing list archive at Nabble.com.

______________________________________________
R-help_at_stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Received on Thu 03 May 2007 - 21:49:03 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Thu 03 May 2007 - 23:31:52 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.