[R-sig-ME] Specification Noninformative Prior with Jeffrey's Method in MCMCglmm function

[Jarrod Hadfield]

Hi,

I think p(sigma^2) = 1/sigma^2 is the limit of the inverse-chisquare 
distribution as nu goes to zero.  This is the default in MCMCglmm.

Cheers,

Jarrod


On 26/03/2017 12:00, Euis Aqmaliyah wrote:
> Hello.
> Now, i'm writting thesis with the topic Small Area Estimation using
> Bayesian Approach.
> Models that i use in estimation are Linear Mixed Model (LMM) and
> Generalized Linear Mixed Model (GLMM) because my response variable is
> Zero-Inflated Data.
> But, i have problem when i fit a linear mixed model. I want specification
> prior for residual variance parameter (sigma^2) is 1/sigma^2 that i get
> using Jeffrey's Method from Normal distribution.
> So, how to specification that prior in MCMCglmm function?
> I hope you can help me.
> Thank you.
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> 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.