Hello I am using the predict method in fGarch. I tried to replicate what I
supposed it was doing when I estimated an ARMA(2,2)+GARCH(1,1) by doing:
(temp is my ouput form garchFit)
fit = temp@fit$par
Nobs = length(temp@data)
# Predecir media
fit["mu"]+sum(fit[2:(2+2-1)]*temp@data[length(temp@data):(length(temp@data
)-2+1)]),0)+
sum(fit[(2+2):(2+2+o2-1)]*temp@residuals[length(temp@residuals
):(length(temp@residuals)-2+1)]),0)
and I got -0.005585084 while the output from the predict method gave me
-0.00752829. Does anybody know what might be going on? am I using the wrong
slots in the output?
Thank you
Felipe Parra
[[alternative HTML version deleted]]