DocumentCode :
3107886
Title :
A rapid-prototyping approach to fuzzy system modeling
Author :
Hsu, Ya-Chen ; Chen, Guanrong
Author_Institution :
Dept. of Electr. & Comput. Eng., Houston Univ., TX, USA
fYear :
1998
fDate :
20-21 Aug 1998
Firstpage :
15
Lastpage :
19
Abstract :
We develop a three-phase design scheme for fuzzy modeling using input-output training data. The first step is to establish an approximate rule base using a clustering method, then the selective rule activation technique is applied to resize the rule base, and finally the parameters are fine-tuned by the backpropagation algorithm. Simulation results show that this scheme generates a good fuzzy model that can successfully predict the system output and the rules selected by the selective rule activation technique are appropriate for generalizing different sets of data
Keywords :
backpropagation; fuzzy systems; generalisation (artificial intelligence); knowledge based systems; modelling; approximate rule base; backpropagation; clustering method; fuzzy system modeling; generalization; input-output training data; parameter tuning; rapid prototyping; selective rule activation technique; simulation; three-phase design scheme; Backpropagation algorithms; Clustering algorithms; Clustering methods; Fuzzy sets; Fuzzy systems; Humans; Input variables; Partitioning algorithms; Predictive models; Training data;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Fuzzy Information Processing Society - NAFIPS, 1998 Conference of the North American
Conference_Location :
Pensacola Beach, FL
Print_ISBN :
0-7803-4453-7
Type :
conf
DOI :
10.1109/NAFIPS.1998.715520
Filename :
715520
Link To Document :
بازگشت