Date:         Sat, 14 Jan 2006 11:28:52 -0800
Reply-To:     dave@AUTOBOX.COM
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         dave@AUTOBOX.COM
Organization: http://groups.google.com
Subject:      Re: Moving average
Comments: To: sas-l@uga.edu
In-Reply-To:  <3cf9f5a80601140849h2a66e3b1gb178c6afb06e8abf@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

j.

Why would you want to

1. Use all of the values rather than a recent set to compute this
average. perhaps you would be better of just using the most receent
three values ?

2. After you decide or use analytical methods to determine how many
recent values to use . Why would you want to equally weight (believe)
the values . For example if you had determined that thrree (3) periods
was the optimal number to use to create the "average" or representative
value
you might find that

3/6, 2/6 and 1/6 was more appropriate than 2/6,2/6,2/6

For more on time series and attempts to characterize or represent see
http://www.autobox.com

dave r
afs