Date:         Sat, 27 Nov 2004 07:09:36 -0500
Reply-To:     Ben <benpub7@YAHOO.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Ben <benpub7@YAHOO.COM>
Subject:      Re: How to determine ........

Art,
I will try it with sufficient loops. the data is for hundred's id, each of
id has dozens' observations. we want to know every id's most recent, most
friquent activity. variable is ordinal numbers. you are right, it is not
exactly a time series, we don't calculate the MVA etc. it is longitudinal
data.

Thanks



On Fri, 26 Nov 2004 18:46:45 -0500, Arthur Tabachneck <art297@NETSCAPE.NET>
wrote:

>Ben,
>
>Time series are not in my areas of expertise but, that aside, you can
>accomplish what you ask by writing a macro with sufficient loops.
>
>Conversely, why the 'mode' as the measure of central tendency, and not
>the 'mean' or 'median'?  In SAS 9 either can be accomplished quite readily
>and, on a running basis, would only require a simple loop and neither
>necessitating that you be concerned about ties.
>
>Art
>---------
>On Fri, 26 Nov 2004 15:03:36 -0500, Ben <benpub7@YAHOO.COM> wrote:
>
>>can we go further?
>>
>>The datasets is a time series, I want to know the running mode of values
of
>>previous 6 to 10 observations.
>>
>>even further ....
>>
>>what if I want to make a decision depends on whether or not the modes is
>tie
>>or not? univariate won't tell it, if I am correct.