Date: Fri, 14 Jul 2000 19:50:00 GMT
Reply-To: Ben Krone <bkrone@MY-DEJA.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Ben Krone <bkrone@MY-DEJA.COM>
Organization: Deja.com - Before you buy.
Subject: Re: Least Squares fit slope
In article <396F6288.7E2861F5@kodak.com>,
Paige Miller <paige.miller@kodak.com> wrote:
> Ben Krone wrote:
> >
> > Hello
> >
> > Am wanting to take a set of production data measurements and
determine
> > the numeric slope for a least squares fit line to determine if a
> > particular location is improving or declining over the selected time
> > period. An Excel equivalent function would be LINEST or SLOPE.
> >
> > Data looks like so
> >
> > LOCATION DATE MEASURE
> > XXX 01Jan2000 4800
> > YYY 01Jan2000 4126
> > XXX 02Jan2000 4765
> > YYY 02Jan2000 4360
> > XXX 03Jan2000 4915
> > .....
> >
> > What I am looking for is a value that tells me for each location if
> > their measure has improved and to what degree.
> >
> > I am sure SAS can do it, I just can't figure out how.
>
> PROC REG will compute slopes for you, assuming your dates have been
> stored as numerics and not characters. You will need a BY statement,
> specifically BY LOCATION; in order to get different slopes for the
> different locations. As far as determining if things have improved,
that
> requires some human judgment which SAS cannot provide (or perhaps, the
> statement of a specific testable hypothesis, which I don't see in your
> description, in which case, SAS might be of help).
>
> --
> Paige Miller
> Eastman Kodak Company
> paige.miller@kodak.com
>
> "It's nothing until I call it!" -- Bill Klem, NL Umpire
> "Those black-eyed peas tasted all right to me" -- Dixie Chicks
>
Paige,
Thanks for the response...That does the trick. Getting the slope was
really what I was after. If the particular measure's desired result is
a higher value over time, I can look for a positive slope. Plus I can
then compare slope values to determine which locations have the higher
rate (slope) of improvement.
Thanks again
Ben
Sent via Deja.com http://www.deja.com/
Before you buy.
| 
