Message-ID: <20140206151957.208841061ad30so4@www.staffmail.ed.ac.uk>
Date: 2014-02-06T15:19:57Z
From: Jarrod Hadfield
Subject: Specify prior for sir-function and starting values in the MCMCglmm packages?
In-Reply-To: <AD0050057515F54084E7D5B93478C8481F7BA0260B@winxbede39.exchange.xchg>
Hi Linus,
Sorry for the delay. The warning
"priors for sir parameters not implemented"
is correct, and so the priors are essentially flat priors, equivalent
to a Gaussian with mean zero and very large variance. The default
prior in MCMCglmm for the fixed is mean zero and a variance of 10^8
and so I would expect them to give very similar answers (unless there
is confounding). For some models/data (for example categorical data
and complete or near-complete separation) then these priors can behave
poorly, but given sir models can only be fitted to Gaussian data in
MCMCglmm, I guess this is not the case.
In the next update (which I realise is taking a long time) I will
implement prior distributions and starting values.
Cheers,
Jarrod
Quoting Linus Holtermann <holtermann at hwwi.org> on Fri, 31 Jan 2014
12:24:56 +0100:
> Dear list members,
>
> unfortunately Jarrod Hadfield is on field work until the end of June
> and cant help me with my problems. Maybe someone on the mailing
> list knows how to specify starting values and a Prior for the
> sir-function in MCMCglmm. Short code examples would be great.
>
> Best regards,
>
>
> Linus Holtermann
> Hamburgisches WeltWirtschaftsInstitut gemeinn?tzige GmbH (HWWI)
> Heimhuder Stra?e 71
> 20148 Hamburg
> Tel +49-(0)40-340576-336
> Fax+49-(0)40-340576-776
> Internet: www.hwwi.org
> Email: holtermann at hwwi.org
>
> AmtsgerichtHamburg HRB 94303
> Gesch?ftsf?hrer: Prof. Dr. Thomas Straubhaar, Gunnar Geyer
> Umsatzsteuer-ID: DE 241849425
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
>
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.