CODEWITHSUNDEEP
rparameterstime-series
megatron666
megatron666

Reputation: 31

fGarch package in R: parameter definitions / model specification missing?

I am currently using the fGarch package in R and so far really like it. However, one thing I find rather confusing is that there does not seem to be any documentation on the meaning of the parameters? Definitions / choice of variables differs throughout literature..

For instance, if I fit a Garch(1,1) model with skewed-student-t innovations, I obtain the following parameter-vector

garchmod  <-  garchFit(~garch(1, 1), data=m.sarima$fit$residuals,cond.dist="sstd",include.skew=FALSE,include.mean=FALSE)
garchmod$params
   mu     omega    alpha1    gamma1     beta1     delta      skew     shape 
0.0000000 4.7040111 0.4401784 0.1000000 0.3471025 2.0000000 1.0000000 3.5497162

I think omega is the mean in the variance euqation, alpha1 the loading on past squared noise, beta1 the loading on past condt. variance.. Anyone familiar with any documentation on this? Any help greatly appreciated.

Best regards megatron

Upvotes: 3

Views: 554

Answers (1)

Kidae Kim
Kidae Kim

Reputation: 499

Your guesses are correct.

Refer to the parameter details of garchSim(pg14) and GarchFitting(pg6) in the following documentation of fGarch.

http://home.agh.edu.pl/~kostrzew/fGarch.pdf

Upvotes: 0

Related Questions