Date:         Fri, 11 Jul 2008 21:58:49 -0700
Reply-To:     KL <Kev.L.Scott@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         KL <Kev.L.Scott@GMAIL.COM>
Organization: http://groups.google.com
Subject:      Re: ARMA in SAS
Comments: To: sas-l@uga.edu
Content-Type: text/plain; charset=ISO-8859-1

On Jul 6, 11:41 am, liuwen...@GMAIL.COM (Wensui Liu) wrote:
> Sorry, Sudip,
> I take back what I replied earlier. arima and model procs are only
> available in ets. besides ets, the only implementation of arma i've
> seen is with iml. but i think it is very doable with stat.
> i will play with it and see if i can come up something.
>
> On Sat, Jul 5, 2008 at 2:06 PM, sudip chatterjee
>
> <sudip.memp...@gmail.com> wrote:
> >   Dear Users,
>
> >  Is there anyway to estimate an ARMA model in SAS /STAT instead SAS/ETS ?
>
> --
> ===============================
> WenSui Liu
> Acquisition Risk, Chase
> Email : wensui.x....@chase.com
> Blog : statcompute.spaces.live.com
> ===============================

You can specify a general multivariate log likelihood function for the
ARMA model using PROC NLMIXED in SAS/STAT. Be focused and patient,
very patient.