Date: Tue, 21 Mar 2006 11:24:19 -0500
Reply-To: Fred <ieaggie2002@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Fred <ieaggie2002@GMAIL.COM>
Subject: Re: What's the proper PROC to check the stationary property?
In-Reply-To: <1115a2b00603210758u35e16b8l98f3a1e4464ecb78@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Thanka a lot, Wensui.
The SAS HELP document says we can use Autocorrelation,
that is, the time series is stationary if autocorrelation coeffiecients
decay quickly.
But this is only a visual evidence. Can we get some quantitative measure
to determine these autocorrelation are really small so that the process is
stationary?
On the other hand, I tried to use PROC ARIMA as you refferred to.
But there are different options, such as ADF test, Phillips test, and RW
test.
They need to specify the option value prior to using the test.
For an ordinary time series, how to determine the option? and how
to say safely that the tested process is stationary givent the test result?
Sorry to bring up such annoying questions.
Again, appreciate your kind help.
Fred
On 3/21/06, Wensui Liu <liuwensui@gmail.com> wrote:
>
> proc arima
>
>
> On 3/21/06, Fred <ieaggie2002@gmail.com> wrote:
> >
> > Hi, All
> >
> > I have a time series and want to use SAS to check if it is
> > a stationary process.
> >
> > Would you please give me a hint which procedure in SAS
> > is the proper one to achieve this task?
> >
> > Thanks a lot for your input.
> >
> > Fred
> >
>
>
>
> --
> WenSui Liu
> (http://statcompute.blogspot.com)
> Senior Decision Support Analyst
> Health Policy and Clinical Effectiveness
> Cincinnati Children Hospital Medical Center
>
| 
