Date:         Sun, 12 Nov 2006 18:32:04 -0800
Reply-To:     Cat <job.alerte@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Cat <job.alerte@GMAIL.COM>
Organization: http://groups.google.com
Subject:      PROC CLUSTER and Distance data set, PROC TREE
Comments: To: sas-l@uga.edu
Content-Type: text/plain; charset="iso-8859-1"

Hi SAS-experts !

I need to perform a hierarchical agglomerative cluster analysis on a
set of categorical variables which were collected in around 130
patients. I planned to used the Chi2 distance.

I'm wondering how Proc cluster deals with these input data sets (type
'Distance'). Since the history of aggglomeration strongly depends on
the distance, and since this distance is not specified in input
parameters of this procedure, I can't understand how SAS can determine,
at each step, which clusters must be gathered. Does anyone knows ?

Second question: I tried to develop a code to produce a(n?) horizontal
tree after clustering data using Chi2 distance between individuals. The
tree used, as abscissa, the R-squared... or something like that. Do you
know if it is possible to put the Chi2 distance instead ? How ?

Third question: Among agglomeration methods proposed by SAS, does one
know whether some are not appropriate for categorical variables ? which
ones ? why ?

Thanks a lot for your support.

Regards,

Catherine.