摘要
前向神经网络的泛逼近性一直是神经网络的研究热点。本文给出了连续模糊函数的定义,依Hausdorff度量,借助模糊值Bernstein多项式关于连续模糊函数的逼近性质,证明了前向网络作为模糊函数泛逼近器的一致逼近性结果,并通过实例给出了逼近性的具体实现过程。
Uniform approximation of feedforward neural networks has always been a hot spot on neural networks＇ research. In this paper, the definition of continuous fuzzy functions is given, according to Hausdorff measure, through the approximation of fuzzy-valued Bernstein polynomials on continuous fuzzy functions, the consistent approximation of feedforward neural network as a universal approximator of fuzzy functions is proved, and an example is shown to explain the realization process of the approximation.
出处
《模糊系统与数学》
CSCD
北大核心
2012年第5期132-137,共6页
Fuzzy Systems and Mathematics
基金
天津市高等学校科技发展基金资助项目(20110519)
