Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!news.moneng.mei.com!uwm.edu!news.alpha.net!news.mathworks.com!news.kei.com!wang!news
From: natalieb@iil.intel.com (Baruch Natalie)
Subject: Clusterization Problem
Organization: Intel Israel (74) Ltd.
Date: Fri, 17 Feb 1995 09:53:41 GMT
Message-ID: <3i1rn5$2uni@ilx018.iil.intel.com>
Sender: news@wang.com
Lines: 26

Hi there, Net-Pals,

I need some help. I need to find a solution to the famous clusterization
problem. We, my team and me, are concentrating on Kohonen's map with some help from the LVQ algorithm. However, we seem to be stuck.

Our goal is to be achieved in three stages:
1. Random group division (Simple Kohonen would do this by natural similarity)
2. Dividing input into K groups (where K is predefined and constant)
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).

   If you can offer a practical solution, maybe using Kohonen, LVQ and 
   another FeedForward net connected, please reply. 


   Thanks in advance,

   Natalie

   e-mail: natalieb@iil.intel.com

**********************************************************************
   THIS IS A PERSONAL MAIL.
   ALL THE MATERIAL IS MY PERSONAL OPINION, AND HAS NOTHING TO DO WITH MY I
   EMPLOYER.
**********************************************************************
