[Rd] garchFit (PR#10698)

Yohan Chalabi chalabi at phys.ethz.ch
Tue Feb 5 12:53:25 CET 2008

>>>> "SYC" == scott.yonker at gmail.com
>>>> on Mon,  4 Feb 2008 23:35:07 +0100 (CET)

   SYC> The garchFit function in the Rmetrics libray cannot estimate
   SYC> ARCH models,
   SYC> meaning that the number of beta terms in the model must be
   SYC> greater than zero.
   SYC> For this function neither p nor q can be set to zero without
   SYC> an error.  The
   SYC> problem lies in both the specification of the objective function
   SYC> and in the
   SYC> function that determines the initial parameter estimates.
   SYC> A simple if statement
   SYC> setting alpha=0 if p=0 and beta=0 if q=0 (I don't remember
   SYC> which is p or q I
   SYC> might have these reversed) in both of these functions solves
   SYC> the problem.

Hi Scott,

Thank you for this information. I will care about it.

But please use the Rmetrics core team list (rmetrics-core at r-project.org) to send
bugs related to Rmetrics.


More information about the R-devel mailing list