摘要
本文提出了一种能进行模糊逻辑计算的Ｈｏｐｆｉｌｅｄ型人工神经元网络（ＦｕｚｙＨＮ），ＦｕｚｚｙＨＮ的神经元对应模式集合中的元素，模式间的模糊相似关系作为联结神经元的权值矩阵被存储在ＦｕｚｚｙＨＮ中。本文对ＦｕｚｙＨＮ的稳定性及模糊聚类功能进行研究，获得了良好的理论分析结果和实验研究结果。本文的研究表明。
This paper proposes a fuzzy Hopfield network (FuzzyHN) with the computational functions of fuzzy logic.The neurons in FuzzyHN corresponds to the elements in a certain pattern set,and the fuzzy relation between the patterns is stored as the weighting matrix in FuzzyHN.The author analyses theoretically the stability and the clustering function of FuzzyHN.The study on FuzzyHN in this paper has shown that FuzzyHN possesses better stability and fuzzy clustering function that is equivalent to fuzzy clustering operation based on fuzzy equivalent relation.FuzzyHN as a computational model of fuzzy logic supplies a new neural network based imlementation for fuzzy pattern recognition.
出处
《模糊系统与数学》
CSCD
1997年第1期88-94,共7页
Fuzzy Systems and Mathematics
基金
航空科学基金
