Structure conclusion identification to improve the fuzzy models accuracy
S. Teghbit, N. Malvache, D. Willaeys · Proceedings of IEEE 5th International Fuzzy Systems · 2002
The subject deals with the accuracy problems in modeling systems, when the process is linguistically described by a set of rules such as "if the input variable is A/sub i/ then the output variable is B/sub j/". The influence of the fuzzy sets structure on the fuzzy model is studied, and an appropriate representation, such as the margin between the real system and the obtained model reaching its minimal value, is proposed.