摘要
基于 J.F.Baldwin等人提出的 mass- assignment理论 ,提出了新的基于 mass- assignment的模糊 CMAC神经网络 ,接着研究了其学习规则 .理论研究结果揭示出 ,此新模糊 CMAC是一个全局逼近器 ,并且具有学习收敛性 .故此新模糊 CMAC有非常重要的应用潜力 .
In this paper, based on the mass-assignment theory proposed by J.F. Baldwin et al., the new mass-assignment-based fuzzy CMAC is presented. Accordingly, its learning rules are also investigated. The theoretical research results reveal that this new mass-assignment-based fuzzy CMAC is a universal approximator, and has its learning convergence. Therefore, this new fuzzy CMAC has very important potentials of applications.
出处
《软件学报》
EI
CSCD
北大核心
2001年第6期816-821,共6页
Journal of Software
基金
国家自然科学基金 No.6 98 430 0 2 &&
关键词
模糊CMAC
学习规则
mass-assignment理论
学习收敛性
神经网络
Approximation theory
Convergence of numerical methods
Fuzzy sets
Iterative methods
Learning systems
Mathematical models
Membership functions
Multilayer neural networks
Probability distributions
Theorem proving
Vectors
