摘要
本文采用广义模糊神经网络实现分段建模的思想．给出了一种广义k－均值聚类算法，该算法能同时确定模糊规则的个数和相应的参数．仿真结果表明该算法是可行和有效的．
This paper presents a generalized fuzzy neural network that can realize the strategy about approximating a function using piecewise models. A generalized k- means algorithm is given. The number and parameters of fuzzy rules can be simultaneously obtained by the algorithm. Computer simulations show that the method is feasible and efficient.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
1998年第1期148-151,共4页
Control Theory & Applications
基金
广东省自然科学基金!960101
国家攀登计划认知科学(神经网络)重大关键项目!B1-131-343
