This paper establishes the approximation error bounds for fuzzy systems with the center-average defuzzifier. Based on these bounds, the approximation accuracy of fuzzy systems under different inference methods is compared, which suggest that the class of fuzzy systems generated by the product inference and the center-average defuzzifier have better approximation accuracy and properties than the class of fuzzy systems generated by the min inference and the center-average defuzzifier. In addition, it is proved by using the obtained approximation bounds that fuzzy systems can represent any linear and multilinear functions.
展开▼