[ADMB Users] mcpin

[Mark Maunder]

In that case would

mymodel -noest -nohess -mcmc N -mcsave M -ainp mymodel.par

Work?



From: Ian Taylor [mailto:ian.taylor at noaa.gov]
Sent: Tuesday, January 31, 2012 4:26 PM
To: Mark Maunder
Cc: Mollie Brooks; ADMB Users
Subject: Re: [ADMB Users] mcpin

I thought the model automatically looked in the directory for the .mcm file or whatever it's called. It works to run
mymodel -noest -nohess -mcmc N -mcsave M
which will start at the initial values rather than those in the .pin file, but it definitely makes use of the correct Hessian matrix info.
-Ian

On Tue, Jan 31, 2012 at 4:09 PM, Mark Maunder <mmaunder at iattc.org<mailto:mmaunder at iattc.org>> wrote:
I think there is, but not sure how to do it.

-----Original Message-----
From: Mollie Brooks [mailto:mbrooks at ufl.edu<mailto:mbrooks at ufl.edu>]
Sent: Tuesday, January 31, 2012 3:12 PM
To: Mark Maunder
Cc: ADMB Users
Subject: Re: [ADMB Users] mcpin

Thanks Mark.
I think I already estimated the variance covariance matrix in the first model run. Is there a way to give it to ADMB?



On 31 Jan 2012, at 5:47 PM, Mark Maunder wrote:

> You also need the variance-covariance matrix for the jumping rule. You could use
>
> xxxx -ainp xxxx.par -mcmc 100000000
>
> To start the model at the parameter estimates (makes it run quicker), estimate the hessian (will take some time), and run the mcmc
>
>
>
>
> -----Original Message-----
> From: users-bounces at admb-project.org<mailto:users-bounces at admb-project.org> [mailto:users-bounces at admb-project.org<mailto:users-bounces at admb-project.org>] On Behalf Of Mollie Brooks
> Sent: Tuesday, January 31, 2012 2:16 PM
> To: ADMB Users
> Subject: [ADMB Users] mcpin
>
> Hi,
> I have already found point estimates for the parameters of my model. Now I'm interested in doing mcmc sampling.
> Can I use the command line option -mcpin to start the mcmc sampling at the values stored in model.par without redoing the maximum likelihood estimation? The manual talks about using -mcpin with values from profiling, but I can't do profiling because I'm interested in the distribution of the variance parameters of random effects.
>
> thanks,
> Mollie
>
> Mollie Brooks
> Ph.D. Candidate
> NSF IGERT Fellow
> Biology Department
> University of Florida
> mbrooks at ufl.edu<mailto:mbrooks at ufl.edu>
> http://people.biology.ufl.edu/mbrooks
>
>
>
>
> _______________________________________________
> Users mailing list
> Users at admb-project.org<mailto:Users at admb-project.org>
> http://lists.admb-project.org/mailman/listinfo/users
>

_______________________________________________
Users mailing list
Users at admb-project.org<mailto:Users at admb-project.org>
http://lists.admb-project.org/mailman/listinfo/users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.admb-project.org/pipermail/users/attachments/20120201/eb0e324d/attachment.html>