Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!Germany.EU.net!EU.net!sun4nl!news.nic.surfnet.nl!spider.research.ptt.nl!gerwen.research.ptt.nl!M.J.A.M.vanGerwen
From: M.J.A.M.vanGerwen@research.ptt.nl (Emile van Gerwen)
Subject: Certainty Factors and the combination of evidence
Message-ID: <M.J.A.M.vanGerwen.31.000928D2@research.ptt.nl>
Sender: usenet@news.research.ptt.nl (USENET News System)
Nntp-Posting-Host: gerwen.research.ptt.nl
Organization: PTT Research
X-Newsreader: Trumpet for Windows [Version 1.0 Rev A]
Date: Wed, 11 Jan 1995 08:10:30 GMT
Lines: 28



Recently I decided to use the Certainty Factor model as a way to handle 
uncertainty in my application. Browsing the literature I noticed that the rule 
that deals with the combination of two pieces of evidence for the same 
hypothesis is defined as:

CFresult = (CF1 + CF2) / (1 - min{abs(CF1),abs(CF2)})    if  -1 < CF1*CF2 < 0

I was wondering why the rule is defined as above. Intuitively I would have 
defined the rule simply as:

CFresult = CF1 + CF2    if -1 <= CF1*CF2 < 0

The only justification for the first rule I found in the literature is 
"pragmatic reasons". 

Can someone help me in understanding this rule??

Thank you very much.
Emile van Gerwen.

--------------------------------------------------------------------------
--  Emile van Gerwen             
--  PTT Research                            P.O.Box 421
--  Tel: +31-70-3326126                     2260 AK Leidschendam
--  Fax: +31-70-3326477                     The Netherlands
--------------------------------------------------------------------------
