Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!cs.utexas.edu!utnut!cannon.ecf!srodula
From: srodula@ecf.toronto.edu (SRODULA  EWA ANNA)
Subject: Q: References for Levenberg-Marquardt algorithm
X-Nntp-Posting-Host: spark21.ecf
Message-ID: <D6JCuu.3wu@ecf.toronto.edu>
Sender: news@ecf.toronto.edu (News Administrator)
Organization: University of Toronto, Engineering Computing Facility
Date: Wed, 5 Apr 1995 00:18:30 GMT
Lines: 17

Hello everyone,

I have been exploring the new (version2.0) Neural Net Toolbox for
MATLAB and tried the Levenberg-Marquardt algorithm on some
(offline) adaptive filtering problems. It seems to do a much
better job than Conjugate Gradient based learning (source code
from the book Neural Network Recipes in C++ by Timothy Masters).
Unfortunately, the User Manual does not suggest any relevant
references on Levenberg-Marquardt method. Is it yet another
optimization scheme like conjugate gradient? What advantages
does it hold over conj.grad method?
Could someone kindly suggest some good references on this topic?
I will compile and post the responses.

Thank you.

-Thas (thas@ibme.utoronto.ca)
