[R] multivariate simulation

[Thomas Lumley]

On Wed, 23 Jan 2002, Bernard Colin wrote:

> To whom it may concern,
>
> I try to simulate a non-normal multivariate distribution. The MASS package
> allows by mean of "mvrnorm" command to perform a multivariate normal
> simulation. Is there an equivalent command for an arbitrary multivariate
> distribution available in the R-language?
>

No. An `arbitrary multivariate distribution' covers a whole lot of
possibilities.

It would be fairly easy to write a function that linearly transformed an
arbitrary starting distribution to have arbitrary mean and variance in the
same way that mvrnorm() does, but that is usually not desirable.

It is unusual (? unique) to the Normal that this procedure produces a
random vector that is marginally and conditionally Normal.  An analogous
mvrgamma would produce variables that are not Gamma distributed and an
mvrpois wouldn't even produce integers.

There are many other ways of producing multivariate versions of univariate
distributions and most of them are `natural' for only a few cases.

If you could be more specific about what you want someone might be able to
help, but most of these methods aren't programmed in R. Jim Lindsey's
packages have a few.

	-thomas

Thomas Lumley			Asst. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._