Fuzzy linguistic modifiers in expert systems


Subject: Fuzzy linguistic modifiers in expert systems
From: Otto Cordero (ocordero@cti.espol.edu.ec)
Date: Wed Oct 04 2000 - 12:18:05 MET DST


This is a multi-part message in MIME format.

------=_NextPart_000_005B_01C02C92.142B4000
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
        charset="iso-8859-1"

Hi, i need to define rules for an expert system with fuzzy linguistic =
terms, the problem is that we dont know what the concepts involved in =
the rule are, since we have developed a technique to allow the expert to =
register his or her knowledge, all concepts are symbolic, for example:

if we have CONCEPT1 then we have a lot of of CONCEPT2

CONCEPT1 and CONCEPT2 are symbolic terms that define some knowledge of =
the expert. This rule is expresed in the form of a cognitive map. with =
an arrow from CONCEPT1 to CONCEPT2.=20
=20
How can i cuantify the presence of C2?. If we would know what C2 is we =
could define a fuzzy set, for example: VERY HEAVY=3D{ 0.6/80kg, =
0.8/90kg, 0.95/100kg }
but as i said C1 and C2 are symbolic. What we are doing now is using =
closed intervals between 0 and 1 to define the linguistic modifiers, for =
ex: high=3D0.7<x<=3D1.0; medium=3D0.4<x<=3D0.7 and so on.....but this =
goes in contrast with the escence of fuzzy logic and doesn't let us take =
advantage of fuzzy techniques.=20
with our aproach the rule would be reduced to:
 if C1 then 0.8 of C2, this is what i need to improve defining really =
fuzzy sets instead of discrete weights.
I am asking you for suggestions in these theme. Or else if you know a =
paper that covers this issue let me know please.
Thank you
Otto Cordero
CTI - ESPOL=20

------=_NextPart_000_005B_01C02C92.142B4000
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
        charset="iso-8859-1"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">

Hi, i need to define rules for an = expert system=20 with fuzzy linguistic terms, the problem is that we dont know what = the=20 concepts involved in the rule are, since we have developed = a technique to=20 allow the expert to register his or her knowledge, all concepts are=20 symbolic, for example:
 
if we have CONCEPT1 then we = have a lot=20 of of CONCEPT2
 
CONCEPT1 and CONCEPT2 are symbolic = terms that=20 define some knowledge of the expert. This rule is expresed in the = form of a=20 cognitive map. with an arrow from CONCEPT1 to CONCEPT2.
 
How can i cuantify the presence of = C2?. If we=20 would know what C2 is we could define a fuzzy set, for example: VERY = HEAVY=3D{=20 0.6/80kg,  0.8/90kg,  0.95/100kg }
but as i said C1 and C2 are symbolic. = What we are=20 doing now is using closed intervals between 0 and 1 to define the=20 linguistic modifiers, for ex: high=3D0.7<x<=3D1.0; = medium=3D0.4<x<=3D0.7 and=20 so on.....but this goes in contrast with the escence of fuzzy logic and = doesn't=20 let us take advantage of fuzzy techniques.
with our aproach the rule would be = reduced=20 to:
 if C1 then 0.8 of C2, this is = what i need to=20 improve defining really fuzzy sets instead of discrete = weights.
I am asking you for suggestions in = these theme. Or=20 else if you know a paper that covers this issue let me know=20 please.
Thank you
Otto Cordero
CTI - ESPOL 
 
------=_NextPart_000_005B_01C02C92.142B4000-- ############################################################################ 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 : Wed Oct 04 2000 - 12:18:34 MET DST