摘要
利用模糊理论和BP网络相结合组成的模糊神经网络系统,能够克服单独使用BP网络的局限性,但采用模糊神经网络解决实际问题时存在的难点是如何确定其结构模型.采用模糊聚类分析法对学习样本进行特征抽取,同时综合考虑样本的输入、输出信息,以确定隐层节点数和网络结构.通过对实验数据的分析,提出的模糊神经网络模型取得了理想的识别效果.
The fuzzy neural network system, which is composed of fuzzy theory and BP network, can overcome the limitation of using BP network alone. But the difficulty of using fuzzy neural network to solve practical problems is how to determine its structure model. In this paper, the fuzzy clustering analysis method is used to extract the characteristics of learning samples. At the same time, the input and output information of the sample are taken into consideration to determine the number of nodes and the network structure of the hidden layer. The application of the actual data shows that the fuzzy neural network model of this paper has achieved an ideal recognition effect.
作者
钮永莉
陈晖
魏光杏
NIU Yong-li;CHEN Hui;WEI Guang-xing(Information Engineering Department,Chuzhou Vocational and Technical College,Chuzhou 239000,Anhui;Mathematics and Computer Science Institute,Fuzhou University,Fuzhou 350019,Fujian,China)
出处
《韶关学院学报》
2018年第6期21-24,共4页
Journal of Shaoguan University
基金
安徽省高校自然科学重点项目(KJ2015A402)