Newsgroups: comp.ai.neural-nets
From: jimmy@ecowar.demon.co.uk (Jimmy Shadbolt)
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!news.sprintlink.net!peernews.demon.co.uk!ecowar.demon.co.uk!jimmy
Subject: Re: Clusterization Problem 
Distribution: world
References: <3ic6de$6mp@ninurta.fer.uni-lj.si>
Organization: Econostat
Reply-To: jimmy@ecowar.demon.co.uk
X-Newsreader: Simple NEWS 1.90 (ka9q DIS 1.21)
Lines: 19
X-Posting-Host: ecowar.demon.co.uk
Date: Thu, 23 Feb 1995 13:50:15 +0000
Message-ID: <793547415snz@ecowar.demon.co.uk>
Sender: usenet@demon.co.uk

In article <3ic6de$6mp@ninurta.fer.uni-lj.si> simond@kette.fer.uni-lj.si writes:

>natalieb@iil.intel.com writes:
>
..
>>3. Dividing into K groups (clusters) and trying to keep the division unified 
>>   (number of inputs per cluster should be about the same for all clusters).
>
>The problem is that most of the conventional VQ methods partition the feature
> space
>into separate regions according to some distortion measure regardless of the 
>feature vectors probability distributions
..
        Any ideas what to do if data are nonstationary?

        Cheers

        Drago

