He guys,
I'm trying to fit an arma(1,1)-garch(1,1) model with gaussian innovations, by manually implementing the likelihoodfunction, and subsequently using maxLik() to find the required estimates of the parameters. I know there is a garchFit()-command that does the same, and I use this command to check whether what I did manually is correct. However, I should use the same initial values for mu_0 and sigma_0^2 otherwise the comparison is meaningless. The problem is that I don't know what initial values are used in the garchFit()-function. I can't find it using the help(garchFit)-command, neither does google provide an answer.
Does anyone of you guys know what initial values are used when one fits a arma(1,1)-garch(1,1) model with gaussian innovations by means of garchFit()?
Kind regards,
BabaSyb
[[alternative HTML version deleted]]