Re: Fitting Multivariate Distributions


Mike Aguilar wrote:
Does anyone have advice on how to fit a t-copula?

For that matter, does anyone know how to fit a simple multivariate
(t) distribution via maximum likelihood?

The context of my project is Risk Management. I'm trying to fit a
t-copula for 10 assets. I can create the copula and generate the
random variates via Sklar's Thrm.

I can use the estimates for Kendall's tau to reduce my estimation
problem down to a single parameter - the degrees of freedom. I even
have the theoretical log likelihood function for the copula. But, I
have no idea how to estimate it.

Got to the CSSM archives in Google Groups and search on
"copula". If that doesn't help, expand the search to other
newsgroups like sci.stat.*.

Hope this helps.

Greg

.

Relevant Pages

  • Fitting Multivariate Distributions
    ... Does anyone have advice on how to fit a t-copula? ... does anyone know how to fit a simple multivariate ... distribution via maximum likelihood? ... have the theoretical log likelihood function for the copula. ...
    (comp.soft-sys.matlab)
  • Re: fitting multivariate t-distribution
    ... My immediate question is to fit a MVT for my four ... I must confess I'm not too familiar with fitting general copula ...
    (comp.soft-sys.matlab)
  • Re: How does the Distribution Fitting Tool compute the t-location scale distribution?
    ... fitting a distribution to it and generating random vectors based on ... To fit a copula, you'd first transform each variable to the Uscale, then fit to the result. ... Ditto all of the above for a MVT, except that the MLEs are a bit harder. ...
    (comp.soft-sys.matlab)
  • Re: different MLE gives different answers????
    ... What is the difference between goodness of fit and maximum ... Maximum likelihood is one of the methods that is ... use different maximum likelihood equations in order to find the best ...
    (sci.stat.math)
  • Re: fitting a Gamma cdf to my data
    ... the following demo contains a section showing one way to fit a gamma ... Then you can use gamfit to get the maximum likelihood estimates. ... mimimize the negative log likelihood weighted by the frequencies. ...
    (comp.soft-sys.matlab)