论文部分内容阅读
本文基于模糊Hopfield网络研究提出了一种神经元模糊识别系统(Neuro-Fuzzy Recognition System,简称NFR系统,或NFRs)。其核心是(N+1)阶模糊Hopfield网络和NFR聚类核。通过(N+1)阶模糊Hopfield网络中的N阶子网络,对样本模式进行模糊聚类,学习样本模式中隐含的模糊聚类结构知识,形成NFR聚类核。基于NFR聚类核形成的知识结构,(N+1)阶模糊Hopfield网络对由待识别模式和样本模式构成的模式集合进行模糊聚类运算。NFRs可对模式空间的模式进行分类和识别,并依样本模式将其划分为等价类。论文对NFRs的性能进行了理论分析和示例研究,结果显示,NFRs具有良好的特性。
In this paper, we propose a Neuro-Fuzzy Recognition System (NFR) based on the fuzzy Hopfield network. At its core are (N + 1) -order fuzzy Hopfield networks and NFR clustering kernels. Through the N-order sub-networks in the (N + 1) th order fuzzy Hopfield network, the sample patterns are clustered with fuzzy clustering to learn the knowledge of fuzzy clustering structure hidden in the sample patterns and form the NFR cluster kernels. Based on the knowledge structure formed by the NFR cluster kernel, the (N + 1) th order fuzzy Hopfield network is used to perform the fuzzy clustering operation on the set of patterns consisting of the pattern to be identified and the sample pattern. NFRs can classify and recognize patterns in pattern space and classify them into equivalence classes according to the sample pattern. In this paper, the theoretical analysis of the performance of NFRs and example studies show that NFRs have good characteristics.