| Date: | Tue, 29 Jan 2008 10:20:32 -0500 |
|---|
| Reply-To: | Wensui Liu <liuwensui@GMAIL.COM> |
|---|
| Sender: | "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> |
|---|
| From: | Wensui Liu <liuwensui@GMAIL.COM> |
|---|
| Subject: | Re: Lag selection in Forecast |
|---|
|
| In-Reply-To: | <OFD3C12CE8.FF7C3ADB-ON832573DF.004E3BC8-832573DF.004EE53E@serasa.com.br> |
|---|
| Content-Type: | text/plain; charset=ISO-8859-1 |
|---|
you need to look at acf and pacf to identify the model spec.
On Jan 29, 2008 9:21 AM, Ricardo G Silva <ricardosilva@serasa.com.br> wrote:
> Dear Users,
>
> I have a simple time-series regression model with 3 covariates, for which
> I need to compute 6 and 12 months ahead forecasts.
> However, since I'm working with monthly time series, I also would select
> lags to include (from 1 to 12).
> How can I perform this in SAS?
> Note that I want to maximize the forecast power and not the model's
> adjustment statistics.
>
> Thanks
>
> Rick
>
--
===============================
WenSui Liu
Statistical Project Manager
ChoicePoint Precision Marketing
(http://spaces.msn.com/statcompute/blog)
===============================
|
