Re: Defuzzification


Subject: Re: Defuzzification
From: Breezy (weiz@spam-block.sympac.com.au)
Date: Mon Nov 20 2000 - 15:24:09 MET


defuzzification is the conversion of a fuzzy logic answer space
into a crisp answer spot (number or other single valued variable).

I still question if it's a natural fuzzy thing to do but some problems
that fuzzy logic is applied to simply require that crisp outcome.

A continuously adjusting control system can better afford a rougher
single point defuzzification result because over or under projection
is corrected as a result of the next defuzzification and on and on.

A problem requiring a once off definitive single point outcome from a fuzzy
logic system is better served, I believe by a series of defuzzification
processes
fired in parallel.

A test data table passed over the rules, a range of defuzzification
methods fired on each rule & on each input's answer space and the
outcome is monitored for accuracy against the known test outcome.

This way particular defuzzification systems may be rated for accuracy for
particular rules. I do this and each rules accuracy is stretched by the
worst
under or over percentage error.

A fuzzy logic system may then be allowed to output a number of answers
resulting from all of the defuzzifications of all the answer spaces
resulting
from all the rules that fired for a particular input set.

The test data will have established a + & - accuracy for @ answer. So by
combining all the ranges of all the answers a most accurate range can be
determined. This way the space becomes finally narrow & allows an answer
at the mid point + or - an error.

the meaning of life might end up as 66 + or - 6 as an example.

Rob W

fractalized wrote in message <8unuqo$4gl$1@nnrp1.deja.com>...
>Hello,
>
>I did Not understand the singleton defuzzification method. I was hoping
>that someone here might be able to tell me how to defuzzify a function
>using either or both singleton methods.
>
>--
>Thanks
>fractalized
>
>
>Sent via Deja.com http://www.deja.com/
>Before you buy.

############################################################################
This message was posted through the fuzzy mailing list.
(1) To subscribe to this mailing list, send a message body of
"SUB FUZZY-MAIL myFirstName mySurname" to listproc@dbai.tuwien.ac.at
(2) To unsubscribe from this mailing list, send a message body of
"UNSUB FUZZY-MAIL" or "UNSUB FUZZY-MAIL yoursubscription@email.address.com"
to listproc@dbai.tuwien.ac.at
(3) To reach the human who maintains the list, send mail to
fuzzy-owner@dbai.tuwien.ac.at
(4) WWW access and other information on Fuzzy Sets and Logic see
http://www.dbai.tuwien.ac.at/ftp/mlowner/fuzzy-mail.info
(5) WWW archive: http://www.dbai.tuwien.ac.at/marchives/fuzzy-mail/index.html



This archive was generated by hypermail 2b25 : Mon Nov 20 2000 - 15:27:00 MET