optimising garch model

[econenthusiast]

Hi guys,

I am trying to run an egarch(p,q) model for a monthly return series, with some dependent variables in the variance equation. I am iterating this over the time horizon using a rolling window (say 150 times) and at the end of each window/iteration I save the 1-step ahead forecast of the model. I use this to see how, historically, the model has behaved. I have successfully programmed this in eviews for fixed values of p & q - however i am now trying to figure out the optimal paramters of p & q for a particular returns series. Whats the best way to get around doing that?

Thanks.