Page 1 of 1
how can i get the best ARMA model for my variable?
Posted: Mon Jan 10, 2011 1:07 pm
by samuelillo89
hello again, I have been reading "introduction to time series and forecasting" by brockwell and davis and i have seen that the last part of the book is a small guide to a program called ITSM, they say it features an option which can give you the best ARMA model for your variable based on the smallest Akaike info criterion, i was wondering if Eviews features something like that but with the best Schwarz info criterion(or if not, with akaike), could anyone briefly explain to me how to do it?
Thanks again.
Re: how can i get the best ARMA model for my variable?
Posted: Mon Jan 10, 2011 1:19 pm
by EViews Gareth
Re: how can i get the best ARMA model for my variable?
Posted: Tue Jan 11, 2011 11:37 am
by samuelillo89
thanks a lot, by the way, which is the best criterion for choosing an univariate arma model?? is it the Schwarz criterion(the smaller) or the akaike?
Re: how can i get the best ARMA model for my variable?
Posted: Tue Jan 11, 2011 11:39 am
by EViews Gareth
Not sure I can give you an answer for that. Personally I believe neither works very well.
Re: how can i get the best ARMA model for my variable?
Posted: Tue Jan 11, 2011 12:05 pm
by samuelillo89
and which criteria or method would you recomend?
Re: how can i get the best ARMA model for my variable?
Posted: Tue Jan 11, 2011 12:12 pm
by EViews Gareth
I wouldn't.
Re: how can i get the best ARMA model for my variable?
Posted: Wed Jan 12, 2011 4:07 am
by 036910
[how can i estimate an ARMA estimation
Re: how can i get the best ARMA model for my variable?
Posted: Wed Jan 12, 2011 9:02 am
by EViews Gareth
If you look up ARMA in the User Guide, you should get a good description of what to do.