Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!europa.eng.gtefsd.com!gatech!concert!sas!mozart.unx.sas.com!saswss
From: saswss@hotellng.unx.sas.com (Warren Sarle)
Subject: Re: DIPOL92 algorithm
Originator: saswss@hotellng.unx.sas.com
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <D12rF6.CJx@unx.sas.com>
Date: Mon, 19 Dec 1994 20:25:54 GMT
References:  <3cntvg$52c@fitzherbert.pavilion.co.uk>
Nntp-Posting-Host: hotellng.unx.sas.com
Organization: SAS Institute Inc.
Lines: 28


In article <3cntvg$52c@fitzherbert.pavilion.co.uk>, vspace@pavilion.co.uk (vspace) writes:
|> I have recently come across a number of references to a hybrid algorithm
|> called DIPOL92, which apparently combines trad stats and ANNs in
|> an extremely productive way. *However* I cannot find any papers
|> detailing this (rather too-good-to-be-true) super-net.  So...

If it seems to good to be true, it usually is. And DIPOL has rather less
connection to traditional statistical discriminant analysis than do the
usual feedforward neural nets. It looks kinda kludgy to me, but I
haven't tried it.

|> Does anyone have any leads for me? I am looking for the literature
|> references rather than any code, as we like to build things from the
|> ground up in this neck of the woods. (Bullish humour!)

The algorithm is briefly described in:

   Michie, D., Spiegelhalter, D.J. and Taylor, C.C. (1994), _Machine
   Learning, Neural and Statistical Classification_, Ellis Horwood.

but there are no further citations in there that I can see.

-- 

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.
