计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (16): 60-62.DOI: 10.3778/j.issn.1002-8331.2010.16.017

• 研究、探讨 • 上一篇    下一篇

新型模糊神经网络模型

李孝忠1,2,李 秋2,张有伟2   

  1. 1.天津科技大学 计算机科学与信息工程学院,天津 300222
    2.天津大学 系统工程研究所,天津 300072
  • 收稿日期:2008-12-18 修回日期:2009-02-24 出版日期:2010-06-01 发布日期:2010-06-01
  • 通讯作者: 李孝忠

New kind of fuzzy neural network model

LI Xiao-zhong1,2,LI Qiu2,ZHANG You-wei2   

  1. 1.College of Computer Science & Information Engineering,Tianjin University of Science & Technology,Tianjin 300222,China
    2.Institute of System Engineering,Tianjin University,Tianjin 300072,China
  • Received:2008-12-18 Revised:2009-02-24 Online:2010-06-01 Published:2010-06-01
  • Contact: LI Xiao-zhong

摘要: 分析了已有模糊神经网络模型结构与学习算法的特点,针对它们收敛速度慢、全局逼近能力差等不足,提出了一种新型的模糊神经网络模型,其在模糊化层实现了隶属函数的合成,且结构简单、推理层只有两个节点。实验结果表明该模型具有收敛速度快、全局逼近能力强的优点,具有一定的实用价值。

关键词: 模糊神经网络, BP算法, 参数优化

Abstract: Model structures and learning algorithms of Fuzzy Neural Network(FNN) introduced before have been analyzed.In order to overcome their slow convergence and bad overall approximation,a new kind of fuzzy neural network model is presented.The model proposed synthesizes membership functions in the fuzzy level and has simple structure.The inference level includes only two nodes.The simulation result shows the model has fast convergence rate and fine overall approximation and is more effective compared with other models.

Key words: Fuzzy Neural Network(FNN), Back-Propagation(BP) algorithm, parameter optimization

中图分类号: