Hello R group,
I'm using the tseries package to forecast variance and I would like to do
the same with correlation.
I can't find any way to do that with the function garch().
for example,
I have a matrix of time series
Date CACIndex SPXIndex DAXIndex NKYIndex
1 05/01/1998 3072.84 977.07 4384.81 14896.1
2 06/01/1998 3037.73 966.58 4352.63 14896.40
3 07/01/1998 3006.73 964.00 4391.54 15028.17
4 08/01/1998 2954.94 956.05 4347.23 15019.18
5 09/01/1998 2919.81 927.69 4236.94 14995.10....
and I want to predict the correlation matrix,
how can I do that?
Thanks for your help
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> Hello R group, > I'm using the tseries package to forecast variance and I would like to do > the same with correlation.[...]> and I want to predict the correlation matrix, > how can I do that?You cannot. A (standard) Garch is univariate, correlation requires a multivariate model. There is a cottage industry of papers on multivariate Garch out there, but the problem is hard (in both a statistical and numerical sense). Tell your employer to buy a commercial risk management package. The better ones include something like this. Dirk -- According to the latest figures, 43% of all signatures are totally worthless. -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._