home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai.fuzzy
- Path: sparky!uunet!gatech!destroyer!sol.ctr.columbia.edu!ira.uka.de!gmd.de!NewsWatcher!user
- From: gebhardt@gmd.de (Friedrich Gebhardt)
- Subject: Fuzzy logic and probability: entropy
- Message-ID: <gebhardt-210193081528@129.26.128.171>
- Followup-To: comp.ai.fuzzy
- Sender: news@gmd.de (USENET News)
- Nntp-Posting-Host: 129.26.128.171
- Organization: GMD
- Date: Thu, 21 Jan 1993 07:22:07 GMT
- Lines: 16
-
- Kosko argues somehow like this (I know I exaggerate):
-
- 1. There is something in probability theory called entropy.
- 2. There is something in fuzzy set theory called entropy which
- 2.1 has certain similarity to the probabilistic entropy,
- 2.2 can be derived immediately from first principles of max-min.
- 3. Therefore fuzzy set theory starts where Western logic ends.
-
- And the same with Bayes½ theorem.
-
- Incidentally, because his entropy does not work with t-norms other than
- min, min is the only t-norm that makes sense (p. 277).
-
- Friedrich Gebhardt
- Gesellschaft fuer Mathematik und Datenverarbeitung mbH (GMD)
- D-5205 Sankt Augustin 1
-