Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!news.duke.edu!concert!sas!mozart.unx.sas.com!saswss
From: saswss@hotellng.unx.sas.com (Warren Sarle)
Subject: Re: Masters' Kohonen (SOM) net fails to train.  Why?
Originator: saswss@hotellng.unx.sas.com
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <CyCKw1.L12@unx.sas.com>
Date: Thu, 27 Oct 1994 20:00:01 GMT
References:  <1705A134C0.KPARSONS@kentvm.kent.edu>
Nntp-Posting-Host: hotellng.unx.sas.com
Organization: SAS Institute Inc.
Lines: 29


In article <1705A134C0.KPARSONS@kentvm.kent.edu>, KPARSONS@kentvm.kent.edu writes:
|> ...
|> I want to determine if these responses can be classified
|> into 8 bins (fewer than 8 would not train) using a Kohonen
|> Neural Network?  I am now using the Kohonen network from
|> the `Practical Neural Network Recipes in C++' by Timothy
|> Masters.  I tried SNNS, but couldn't see how to make it
|> learn automatically; it seems you have to repeatedly
|> select randomize weights and click the all button.  I also
|> tried Xerion, but could not get it to compile under Linux
|> (spent a day, made good progress; maybe another day it
|> would compile).
|> 1. Has anyone successfully used Timothy Masters' Kohonen
|> network?  The problem I am having is that the training
|> takes for ever.

You would not have these problems if you used widely-available cluster
analysis software such as that in BMDP, CLUSTAN, JMP, NT-SYS, SAS, SPSS,
etc. There are many clustering programs that are efficient, reliable,
and easy-to-use, and will produce results similar to or better than
Kohonen networks or other neural varieties of adaptive vector
quantization.

-- 

Warren S. Sarle       SAS Institute Inc.   The opinions expressed here
saswss@unx.sas.com    SAS Campus Drive     are mine and not necessarily
(919) 677-8000        Cary, NC 27513, USA  those of SAS Institute.
