CLASS-L Archives

June 2004

CLASS-L@LISTS.SUNYSB.EDU
Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
TwoStep clustering method comparison
From:
Jason Niles <[log in to unmask]>
Reply To:
Classification, clustering, and phylogeny estimation
Date:
Wed, 23 Jun 2004 10:37:35 -0400
Content-Type:
text/plain
Parts/Attachments:
text/plain (15 lines) 
At the recent RC33 conference, a paper called "SPSS TwoStep Cluster -- A
First Evaluation" was presented by Johann Bacher,Knut Wenzig,and Melanie
Vogler -- see

http://www.soziologie.wiso.uni-erlangen.de/ws0304/info/rc33/pp-w-abstracts-rc33.pdf

The authors comparison of TwoStep vs. latent class concluded:

"... SPSS TwoStep algorithm realizes the right cluster structure only for 3
of the 12 tested models (25%). The success rate of LatentGOLD and ALMO is
clearly higher and reaches 100% for LatentGOLD resp. 66% for ALMO - in other
words: LatentGOLD predicts the correct number for all analyzed
constellations, ALMO for 8 of the 12 models. LatentGOLD shows the best
performance."

ATOM RSS1 RSS2