Title :
A robust fuzzy CMAC for function approximation
Author :
Horng-Lin Shieh ; Bao, Chin-Yun
Author_Institution :
Dept. of Electr. Eng., St. John´´s Univ., Tamsui, Taiwan
Abstract :
This paper proposes a new robust fuzzy CMAC algorithm for function approximation. The advantages of CMAC neural network are fast learning convergence, capable of mapping nonlinear functions quickly due to its local generalization of weight updating. In this paper, in order to overcome the problems of function approximation for a nonlinear system with noise and outliers, a robust fuzzy clustering method is proposed to greatly mitigate the influence of noise and outliers and then a new CMAC learning process used to learn the nonlinear system´s features for function approximation.
Keywords :
cerebellar model arithmetic computers; function approximation; fuzzy set theory; learning (artificial intelligence); pattern clustering; CMAC learning process; CMAC neural network; function approximation; noise; nonlinear functions; nonlinear system; outliers; robust fuzzy CMAC; robust fuzzy clustering; weight updating; Artificial neural networks; Cybernetics; Equations; Function approximation; Mathematical model; Noise; Robustness; CMAC; Function approximation; Fuzzy; Noises and outliers; Robust;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2010 International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-6526-2
DOI :
10.1109/ICMLC.2010.5580760