【摘要】 文章介绍一种基于神经网络的模糊推理和规则生成方法,该方法在构造网络时能辨识网络结构和参数,且需要很少的先验信息;文章提出一种混合学习方法,该学习方法分两阶段进行学习,第一阶段使用一种改进的竞争学习方法,建立模糊规则。第二阶段,通过梯度下降技术,来优化模糊规则的参数,以达到高性能的模型。学习后的网络,模糊推理系统的参数融于在网络的拓扑中。文章还给出实验数据。更多还原

【Abstract】 This paper introduces a method of fuzzy inference and rule extraction using a neural network which can identify structure and parameter of network,and need a little priori information.This paper proposes a hybrid learning scheme which includes two phases:at the first phase,rule generation using a modified competitive learning,at the second phase,rule parameter tuning using gradient descent learning.After learning,the network encodes in its topology the essential design parameters of a fuzzy inference system.At the same time ,experimental results are provided.更多还原