[R] GARCH-like

Yohan Chalabi chalabi at phys.ethz.ch
Mon May 26 11:43:12 CEST 2008


>>>> "DU" == "DavidM.UK" <david.merritt at bris.ac.uk>
>>>> on Sat, 24 May 2008 10:36:04 -0700 (PDT)


   DU> Renato,
   DU> 
   DU> As you may know the two routines for fitting GARCH models
   DU> are garchFit
   DU> [fGarch] and garch [tseries]. The RMetrics [who are behind
   DU> the fGarch
   DU> library] team have usefully provided a white paper on Garch
   DU> modeling
   DU> available from their website, which includes a simple example
   DU> of how to fit
   DU> a Garch(1,1) model in R. You should be able to manipulate
   DU> that code to fit
   DU> all manor of GARCH variants.
   DU> 
   DU> I don't think either the tseries or fGarch garch routines
   DU> are actually
   DU> implemented in R code but rather they call Fortran /
   DU> C routines.
   DU> 

FYI, fGarch has both implementations : R code and Fortran subroutines!

   DU> You may also wish to consult the product documentation page
   DU> for the GARCH
   DU> toolbox in MATLAB which is available online - it provides
   DU> many useful tips
   DU> for Garch estimation - including "good initial parameter
   DU> estimates"
   DU> 
   DU> Best
   DU> 
   DU> David M


-- 
PhD student
Swiss Federal Institute of Technology
Zurich

www.ethz.ch
www.rmetrics.org

NOTE:
Rmetrics Workshop: http://www.rmetrics.org/meielisalp.htm
June 29th - July 3rd Meielisalp, Lake Thune, Switzerland



More information about the R-help mailing list