Thread Subject: How to generate non-normal correlated random vars

Hi,

In the past, I used choleskey with correlation matrix to generate correlated random numbers for 12 distributions.

Now I have historical data that fit into arbitrary distributions (similar to generalized pareto). How do I generate correlated random numbers? what information do I need?

Thanks,
dandas

dandas wrote:

> Now I have historical data that fit into arbitrary distributions (similar to generalized pareto). How do I generate correlated random numbers? what information do I need?

One possibility is to use the functions in the Statistics Toolbox for copulas. This demo

<http://www.mathworks.com/products/statistics/demos.html?file=/products/demos/shipping/stats/copulademo.html>

gives the basic ideas behind those functions, while this documentation

<http://www.mathworks.com/access/helpdesk/help/toolbox/stats/brklrj3.html#bqttfgl-1>

goes a little deeper into the details of the functions themselves.

This demo

<http://www.mathworks.com/products/statistics/demos.html?file=/products/demos/shipping/stats/nonparametricCDFdemo.html>

might also be of interest.

Hope this helps.

Thanks a lot for saving me a lot of time.

Peter Perkins <Peter.Perkins@MathRemoveThisWorks.com> wrote in message <hdeju4$ism$1@fred.mathworks.com>...
> dandas wrote:
>
> > Now I have historical data that fit into arbitrary distributions (similar to generalized pareto). How do I generate correlated random numbers? what information do I need?
>
> One possibility is to use the functions in the Statistics Toolbox for copulas. This demo
>
> <http://www.mathworks.com/products/statistics/demos.html?file=/products/demos/shipping/stats/copulademo.html>
>
> gives the basic ideas behind those functions, while this documentation
>
> <http://www.mathworks.com/access/helpdesk/help/toolbox/stats/brklrj3.html#bqttfgl-1>
>
> goes a little deeper into the details of the functions themselves.
>
> This demo
>
> <http://www.mathworks.com/products/statistics/demos.html?file=/products/demos/shipping/stats/nonparametricCDFdemo.html>
>
> might also be of interest.
>
> Hope this helps.