Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!hookup!swrinde!pipex!uknet!festival!ajmy
From: ajmy@festival.ed.ac.uk (A Myles)
Subject: Re: "Trading... - Oh... really?" Part II (last)
References: <83591.arcon@dial.illinois.net>   <goldenjb-261094142042@129.59.170.62> <1994Oct27.115504.26956@oxvaxd> <goldenjb-281094120415@129.59.170.62>
Message-ID: <CyIH25.6Au@festival.ed.ac.uk>
Organization: Edinburgh University
Date: Mon, 31 Oct 1994 00:22:52 GMT
Lines: 31

goldenjb@ctrvax.vanderbilt.edu (jim golden) writes:

>Ah, but you see you do have statistical knowledge which enabled you to
>build a training set of some value.  And you got a net that works.  How
>often does it work?  What architecture did you use?  Why?  How many hidden
>layers?  Why?  What function were you approximating?  How fast does it run?
> How well does the net genralize?  Prove it.  I understand the
>non-disclosure part, the questions are rhetorical.  But if you have
>knowledge of statistics, why did you use a NN at all?  Why not some easily
>implementable statistical method?  My major gripe is that people who use
>NNs without understanding how a network works or why a particular
>architecture was chosen do not get an answer they can justify.  Neural
>networks are very combinatorial entities and every parameter effects
>performance, from momentum to number of layers to squashing function.  I
>believe you should be able to justify every architectural parameter.  

Pardon for stepping in here, I've just noticed this thread today.

Anyway, I've seen a transcribed DTI (Dept. of Trade and Industry) report on
the use of NNs in "the real world" in which a certain large UK chain of
chemist's shops (drug stores) use neural nets (of an unspecified type) to
predict performance.

Most amusing claimed advantage was that no interpretation of the
results is required as with regression analysis (which must
clearly be a different kind of parametric model fitting and
analysis). No disadvantages listed.

Does Dennis Cook know he is wasting his time? :)

	Andy
