Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!caip!elbereth!rutgers!sri-spam!sri-unix!hplabs!sdcrdcf!ucla-cs!zeus!berke
From: berke@zeus.cs.ucla.edu (Peter Berke)
Newsgroups: sci.math.stat
Subject: Connectionism = Statistics?
Message-ID: <2198@curly.ucla-cs.ARPA>
Date: Sat, 18-Oct-86 18:21:23 EDT
Article-I.D.: curly.2198
Posted: Sat Oct 18 18:21:23 1986
Date-Received: Tue, 21-Oct-86 06:44:15 EDT
Reply-To: berke@zeus (Peter Berke)
Distribution: net.math.stat
Organization: UCLA Computer Science Department
Lines: 18

Many "connectionist" models remind me of methods to get statistical
programs (e.g., nonlinear regression, structural relations) to 
converge.  If you know of any analysis or comparisons of these two
classes of research I would like very much to be made aware of them.
My own statistical background is what I might call comfortable to
extensive for a programmer, but minimal for a statistician, so I
am unable to produce such analyses.  I should be able to read them
though, so please, fill me in.

Thank you,
Peter Berke

(please respond by mail, and if I will summarize to net if others would
like)

Hey, also, if you don't know of any such analysis, but would like to
collaborate on some and you have a great knowledge of statistics (for
a statistician), let me know that too.  Thanks, P.