Xref: utzoo comp.graphics:2306 comp.sources.wanted:3916
Path: utzoo!mnetor!uunet!husc6!mailrus!nrl-cmf!ames!pasteur!ucbvax!hplabs!cae780!leadsv!esl!rww
From: rww@esl.UUCP (Richard W. Webb)
Newsgroups: comp.graphics,comp.sources.wanted
Subject: Clustering Algorithms
Message-ID: <655@esl.UUCP>
Date: 20 Apr 88 20:50:27 GMT
References: <615@ubu.warwick.UUCP> <6196@cit-vax.Caltech.Edu>
Reply-To: rww@esl.UUCP (Richard W. Webb)
Distribution: comp.graphics,comp.sources.wanted
Organization: ESL, Inc., Sunnyvale, CA.
Lines: 24
Keywords: clustering, ISODATA

Hello,
	I am working on a personal project that needs to be able to
    automatically find clusters in a set of data points in N dimensions.
    This will be used to iteratively refine the interesting areas in
    certain Fractal objects.  I have also considered using it in a
    hierarchical rendering system.  Given a set of M points, I would
    like any information relevant to the problem of associating points
    with one local adjacent.

	One method I have looked at is to make a Minimum spanning tree
    for the points and throwing away the longest branches until I have
    nice clusters.  This method does work nicely, but has to have some
    arbitrary hueristics to decide when to stop.

	I have heard of a program called ISODATA that does some clustering,
    but I don't know how to get it or what really does.

	Any help or SOURCE :-) would be greatly appreciated.

-- 
Richard W. Webb                ecvax!decwrl!borealis!\
ESL Incorporated                      sdcsvax!seismo!- ames!esl!rww
ARPA: rww%esl@ames.ARPA                ucbcad!ucbvax!/     /
SMAIL: rww@esl.ESL.COM                       ihnp4!lll-lcc!