[R] Does R have an inverse wishart distribution?

Ben Bolker ben at zoo.ufl.edu
Tue Apr 1 15:13:56 CEST 2003


  [Forwarding to R-list as well just so the answer ends up archived 
there.]

  Actually, now that I look again, there's actually a package already on  
CRAN (MCMCpack) that looks like it will do what you want -- 
has random draws and PDF for wishart and inverse-wishart.  It might have 
other useful components as well.
   The only thing I might conceivably have that would go beyond this is 
that the most common algorithm for generating random wishart distributions 
requires that the degrees of freedom be greater than the size (number of 
rows/columns) of the variance-covariance matrix; I coded an alternative 
algorithm without that restriction (from a Springer book by Gentle).

  Let me know if you need it, good luck.

  Ben



On Mon, 31 Mar 2003, Kevin Karty wrote:

> 
> I do in fact mean to sample from it.  Pretty much a basic MCMC using a
> Metropolis Hastings algorithm.  I was unable to find any reference to
> either of these in the help docs.  In particular, am writing a
> hierarchical bayesian multinomial logit application with a mult. Normal
> distribution of coefficients at higher level, so if you know of any
> canned algorithms, that would work as well.  Otherwise, I might be able
> to do some QA work for you. :)  
> 
> 
> -Kevin
> 
> 
> -----Original Message-----
> From: Ben Bolker [mailto:ben at zoo.ufl.edu] 
> Sent: Monday, March 31, 2003 3:25 PM
> To: Kevin Karty
> Cc: r-help at stat.math.ethz.ch
> Subject: Re: [R] Does R have an inverse wishart distribution?
> 
> 
> 
>   Do you want to simulate from it?
>   If so, I could send you various bits of code (some sent by Bill
> Venables 
> to the S-PLUS help list long ago), which I would eventually mean to 
> package and put up.
>  
>   I don't have any code for calculating probability densities (or 
> cumulative distribution functions, or quantiles).
> 
>   Ben Bolker
> 
> On Mon, 31 Mar 2003, Kevin Karty wrote:
> 
> > 
> > 
> > If so, I've had trouble finding it.  Can anyone help?
> > 
> > ______________________________________________
> > R-help at stat.math.ethz.ch mailing list 
> > https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> > 
> 
> 

-- 
318 Carr Hall                                bolker at zoo.ufl.edu
Zoology Department, University of Florida    http://www.zoo.ufl.edu/bolker
Box 118525                                   (ph)  352-392-5697
Gainesville, FL 32611-8525                   (fax) 352-392-3704



More information about the R-help mailing list