lognormal correlated random numbers with Cholesky?

Discuss anything related to quantitative finance which is not directly connected to a particular book here.

lognormal correlated random numbers with Cholesky?

Postby Tesla » Tue Feb 21, 2012 4:49 pm

Hi guys,

So, I have a correlation matrix and I need to draw lognormal random numbers with correlations expressed in the matrix.

Now, we can draw correlated NORMAL distributed random numbers by applying the Cholesky decomposition and multiplying with the normal random number vektor.



But is this only possible for normal distributed random numbers, or is this method valid for all distributions, say also for Inverse Gauß or Weibull?


If not, can I go along and create uniform random numbers, correlate them and then transform them via the inverse cumulative method to say Inverse Gauß or Weibull?

A reply would be realy, realy, realy appreciated!!

Thanks in advance,

Tesla
Tesla
 
Posts: 3
Joined: Wed Jun 29, 2011 3:48 pm

Re: lognormal correlated random numbers with Cholesky?

Postby mj » Thu Feb 23, 2012 10:51 pm

I think people generally correlate normals and then map them to the desired distribution. This is called
using the Gaussian copula.
mj
Site Admin
 
Posts: 1380
Joined: Fri Jul 27, 2007 7:21 am

Re: lognormal correlated random numbers with Cholesky?

Postby Gregorie » Mon Feb 27, 2012 10:12 am

I actually was about to answer, you beat me to it mj..
Gregorie
 
Posts: 1
Joined: Sun Feb 26, 2012 11:38 pm


Return to General technical discussion

Who is online

Users browsing this forum: No registered users and 1 guest

cron