Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!Germany.EU.net!ieunet!news.tcd.ie!kellyfj
From: kellyfj@tcd.ie (Frank Kelly)
Subject: Re: ART2?
Message-ID: <DAHA0L.9Gn@news.tcd.ie>
Keywords: ART2
Sender: usenet@news.tcd.ie (TCD News System )
Organization: University of Dublin, Trinity College
References: <JFSCHREER.4.2FE59092@BIOLOGY.watstar.uwaterloo.ca> <DAFL7p.Lny@unx.sas.com>
Date: Tue, 20 Jun 1995 15:36:21 GMT
Lines: 36

In <DAFL7p.Lny@unx.sas.com> saswss@hotellng.unx.sas.com (Warren Sarle) writes:


>The ART algorithms produce degenerate results for noisy data:

>   Moore, B. (1988), "ART 1 and Pattern Clustering", in Touretzky, D.,
>   Hinton, G. and Sejnowski, T., eds., _Proceedings of the 1988 Connectionist
>   Models Summer School_, 174-185,  San Mateo, CA: Morgan Kaufmann.
>While this article is about ART1, essentially the same problems occur
>with ART2 and FART.

WRONG!!!!!!
I don't know what article on Fuzzy ART you read. Fuzzy ART  does
solve this problem especially the one regarding category proliferation.
Check out their paper!

Their ARTMAP and Fuzzy ARTMAP architectures are also very good 
(they are classifiers, not clustering systems).


>Various clustering algorithms are superior to NN methods for grouping
>data. For example:

>   Balakrishnan, P.V., Cooper, M.C., Jacob, V.S., and Lewis, P.A. (1994)
>   "A study of the classification capabilities of neural networks using
>   unsupervised learning: A comparison with k-means clustering",
>   Psychometrika, 59, 509-525.

You could also try checking out Neural Networks Vol.8 No.2 which has some theoretical
results regarding Fuzzy ART - how it can learn clusterings on one pass in some cases.

--Frank

= Frank.Kelly@cs.tcd.ie  |  AI group, Dept. of Computer Science,   =
=  Work: +353-1-608 1800  |  Trinity College, Dublin 2. Ireland.    =
=       WWW : http://www.cs.tcd.ie/www/kellyfj/kellyfj.html         =
