R help - Sep 2010 - sample exponential r.v. by MCMC

Dear R users,

I am leaning MCMC sampling, and have a problem while trying to sample
exponential r.v.'s via the following code:

samp <- MCMCmetrop1R(dexp, theta.init=1, rate=2,
                          mcmc=5000, burnin=500,
                          thin=10, verbose=500, logfun=FALSE)

I tried other distribtions such as Normal, Gamma with shape>1, it works
perfectly fine. Can someon tell me why?
Thank you for your help.

Best,
Zhongyi

	[[alternative HTML version deleted]]

Hi Dennis,

Sorry for not being considerate. I should have at least mentioned I was
using MCMCpack.
I have no idea about traceback() though.

I appreciate your suggestion.

Best,
Zhongyi



On Wed, Sep 29, 2010 at 5:00 PM, Dennis Murphy <djmuser@gmail.com> wrote:
> Hi:
>
> It might be helpful to inform the list which package you're using -
> evidently it's MCMCpack, but I had to search to learn that.
> You may also want to include the following information:
>
>
> > samp <- MCMCmetrop1R(dexp, theta.init=1, rate=2,
> +                          mcmc=5000, burnin=500,
> +                          thin=10, verbose=500, logfun=FALSE)
> Error in optim(theta.init, maxfun, control = optim.control, lower >
optim.lower,  :
>   non-finite finite-difference value [1]
> > traceback()
> 2: optim(theta.init, maxfun, control = optim.control, lower = optim.lower,
>        upper = optim.upper, method = optim.method, hessian = TRUE,
>        ...)
> 1: MCMCmetrop1R(dexp, theta.init = 1, rate = 2, mcmc = 5000, burnin = 500,
>        thin = 10, verbose = 500, logfun = FALSE)
>
> along with the results of sessionInfo(). You may also want to cc the
> package maintainer in case this happens to be a bug:
> > maintainer('MCMCpack')
> [1] "Andrew D. Martin <admartin@wustl.edu>"
>
> I don't know enough about the package to help you, but providing this
> information to the list will increase the probability of a successful
> resolution in a shorter period of time.
>
> Cheers,
> Dennis
>
> On Wed, Sep 29, 2010 at 2:19 PM, Zhongyi Yuan
<zhongyi-yuan@uiowa.edu>wrote:
>
>> Dear R users,
>>
>> I am leaning MCMC sampling, and have a problem while trying to sample
>> exponential r.v.'s via the following code:
>>
>> samp <- MCMCmetrop1R(dexp, theta.init=1, rate=2,
>>                          mcmc=5000, burnin=500,
>>                          thin=10, verbose=500, logfun=FALSE)
>>
>> I tried other distribtions such as Normal, Gamma with shape>1, it
works
>> perfectly fine. Can someon tell me why?
>> Thank you for your help.
>>
>> Best,
>> Zhongyi
>>
>>        [[alternative HTML version deleted]]
>>
>> ______________________________________________
>> R-help@r-project.org 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.
>>
>
>
	[[alternative HTML version deleted]]