Date: Tue, 18 Oct 2005 11:10:09 -0400
Reply-To: Talbot Michael Katz <topkatz@MSN.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Talbot Michael Katz <topkatz@MSN.COM>
Subject: Re: ordered logit with time series
Hmmm. Well, you might be in trouble. A more detailed explanation of what
you're trying to do would help. First, let's ignore the Time Series
issue. If you're wondering generally about ordinal logit analysis in SAS,
you can handle that with PROC LOGISTIC or PROC GENMOD in SAS/STAT. There
is a nice chapter in Allison's SAS Press book, "Logistic Regression Using
the SAS System" (ISBN 1-58025-352-0) And one of SAS-L's frequent
contributors, Peter Flom, gave a presentation on this topic at the recent
NESUG meeting (I don't know whether he's made the presentation available
for browsing). If the time series data is just among the predictor
variables, i.e., there may be time-dependent effects on the outcome, then
you may be able to feed them into the model as you would any predictor
variable. But if you're trying to forecast ordinal time series outcomes
(which may have autocorrelation, trending, etc.) then you're in deeper
waters. I don't know whether any of the SAS/ETS procedures are designed
to handle ordinal time series outcomes. You could try to mimic time
series analysis by using previous step values of the outcome as predictors
in your ordered logit model, but it's dicey. But perhaps someone else
here knows some ordinal time series tricks in SAS that I don't. As I
said, you're more likely to get useful information if you provide more
detail. Good luck!
-- TMK --
"The Macro Klutz"
On Tue, 18 Oct 2005 04:09:36 -0700, ankitameister@gmail.com
<ankitameister@GMAIL.COM> wrote:
>i am in serious trouble as i have to do ordered logit analysis with
>time series data
>.what is the module for doing it in SAS ?
>any help will be highly appreciated
>thanks in advance
| 
