摘要
本文在深入分析模糊神经系统的基本结构和学习算法的基础上，指出基于梯度下降的学习算法的不足，并提出一种模糊神经系统的修正学习算法。分析表明，该算法具有能够确保模糊语义和模糊空间一致划分的优点。
In this paper,based on deep analysis on the basic structure and learning algorithm for fuzzy neural system,shortcomings of gradient descent based learning algorithm are pointed out,and a modified learning algorithm for fuzzy neural system is proposed.Analysis shows that the modified algorithm has the advantages of stabilizing linguistic meaning of fuzzy sets and keeping the consistency of fuzzy space partition. [WT5”HZ〗
出处
《系统工程与电子技术》
EI
CSCD
1998年第12期75-79,共5页
Systems Engineering and Electronics
关键词
模糊控制系统
学习理论
神经系统
人工智能
Fuzzy neural system,Modified learning algorithm,Fuzzy linguistic meaning,Consistent partition.
