This is a special case of the multivariate t-distribution, which is
defined as a multivariate normal divided by the square root of an
independent scalar chi squared variate divided by its degrees of
freedom.
In S-Plus it seems best to generate such distributions directly
from their definition. Thus to generate a multivariate Cauchy
one might do something like:
rnorm(n)/rnorm(1) ## n is the dimension
If anyone knows better ways to deal with multivariate
t-distributions, I would appreciate the information..
Andrey Feuerverger / Statistics / U of Toronto / andrey@utstat.toronto.edu
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news