[R] Parameterization of Inverse Wishart distribution available in MCMCpack and bayesm libraries

[peter dalgaard]

On Mar 2, 2012, at 05:55 , ilai wrote:

> What do you make of the following from ?riwish
> "
>   riwish(v, S)
> <snip>
>       v: Degrees of freedom (scalar).
> "
> does a m/2 parameterization yield a scalar for, say, 3 dof ?


Er, yes.... (scalar does not imply integer)

As a general matter:

1. This is the Open Source world, you can read the actual function and see what it does. It might even say so in the comments.

2. You can investigate empirically -- the moments are known as a function of the parameters (check Wikipedia), so how about simulating a few thousand matrices and looking at the means and variances. 

(I don't do IW on a daily basis, but AFAICT, the two parametrizations have roughly the same mean, but a factor of two between variances, so it should be fairly easy to spot whether it is one or the other.)

-- 
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com