Re: Simulating Correlated non-Normal Random Variables
- To: mathgroup at smc.vnet.net
- Subject: [mg32641] Re: Simulating Correlated non-Normal Random Variables
- From: Rodney Sparapani <rsparapa at mcw.edu>
- Date: Fri, 1 Feb 2002 02:02:56 -0500 (EST)
- Organization: Medical College of Wisconsin, Milwaukee
- References: <a3apad$49d$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
That transformation trick really only works for multivariate normal, although there is a similar trick for multivariate t. For other distributions there are other tricks which you can find in texts on this subject. But, there is a general solution for most of these problems: Gibbs sampling. However, I haven't seen an implementation of that in Mathematica, but I am interested on working on it. -- Rodney Sparapani Medical College of Wisconsin Sr. Biostatistician Patient Care & Outcomes Research (PCOR) rsparapa at mcw.edu http://www.mcw.edu/pcor Was 'Name That Tune' rigged? WWLD -- What Would Lombardi Do