Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!yeshua.marcam.com!charnel.ecst.csuchico.edu!olivea!decwrl!svc.portal.com!sdd.hp.com!hp-pcd!hpcvusn!nathanm
From: nathanm@cv.hp.com (Nathan K. Meyers)
Subject: Re: stock prediction : (
Message-ID: <1994Oct19.222340.27526@hpcvusn.cv.hp.com>
Sender: nobody@hpcvusn.cv.hp.com (Nobody - UID must be 99999)
Nntp-Posting-Host: hpcvusd.cv.hp.com
Organization: Hewlett-Packard Company, Corvallis, Oregon USA
X-Newsreader: TIN [version 1.1.9 PL6]
References: <1994Oct7.123142.10810@gdstech.grumman.com>
Date: Wed, 19 Oct 1994 22:23:40 GMT
Lines: 18

John Whitson KB2GNC (jwhitson@gdstech.grumman.com) wrote:
: Although this is could be a tangent,  has anyone tried to predict trends
: based upon fixed known cycles?  For example,  some interesting results might
: be found based upon a four-year cycle, otherwise known as one presidential
: term.  
: 
: Of course,  what my broker told me was that the moment a trend is identified,
: there is usually an immediate counter-trend already developed but unnoticed.
: (Sounds almost sinusoidal; anyone ever do spectrum analysis??)

There are very few analytical techniques in the world that haven't been
applied to the stock market... including identification of several
historical cycles of different length (yes, the presidential term is one
of them).  The topic is called "technical analysis" -- books about it
abound, if you're interested in learning more.

Nathan Meyers
nathanm@cv.hp.com
