CAN APPROXIMATE REASONING BE CONSISTENT

Authors
BUCKLEY JJ HAYASHI Y
Citation
Jj. Buckley et Y. Hayashi, CAN APPROXIMATE REASONING BE CONSISTENT, Fuzzy sets and systems, 65(1), 1994, pp. 13-18
Citations number
29
Categorie Soggetti
Computer Sciences, Special Topics","System Science",Mathematics,"Statistic & Probability",Mathematics,"Computer Science Theory & Methods
Journal title
Fuzzy sets and systems → ACNP
ISSN journal
01650114
Volume
65
Issue
1
Year of publication
1994
Pages
13 - 18
Database
ISI
SICI code
0165-0114(1994)65:1<13:CARBC>2.0.ZU;2-X
Abstract
A fuzzy expert system, with one block of rules, is said to be consiste nt if when the input data exactly matches the antecedent of a certain rule, then the final conclusion of the fuzzy expert system exactly mat ches the consequence of that rule. We investigate the two types of app roximate reasoning: (1) first infer then aggregate (FITA), and (2) fir st aggregate then infer (FATI). We argue that FITA can be consistent b ut FATI is, in general, not consistent.