计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (4): 64-66.DOI: 10.3778/j.issn.1002-8331.2009.04.018

• 研究、探讨 • 上一篇    下一篇

对学习矢量量化神经网络中“死”点问题的研究

冯乃勤,南书坡,郭战杰   

  1. 河南师范大学 计算机与信息技术学院,河南 新乡 453007
  • 收稿日期:2008-01-11 修回日期:2008-05-05 出版日期:2009-02-01 发布日期:2009-02-01
  • 通讯作者: 冯乃勤

Research of “blind” spot in LVQ network

FENG Nai-qin,NAN Shu-po,GUO Zhan-jie   

  1. Faculty of Computer & Information Science,Henan Normal University,Xinxiang,Henan 453007,China
  • Received:2008-01-11 Revised:2008-05-05 Online:2009-02-01 Published:2009-02-01
  • Contact: FENG Nai-qin

摘要: 竞争型神经网络已经在模式识别、分类等方面得到了广泛的应用,与传统的聚类方法相比具有巨大优势,但是在许多方面还存在不足,需要进一步完善。在Kohonen提出的学习矢量量化网络(Learning Vector Quantization Network,LVQ)的基础上,引入阈值学习规则,较好地解决了该类网络中遇到“死”点时训练误差偏大的问题,最后通过Matlab编程实现。

关键词: 学习矢量量化网络, 阈值, 死点

Abstract: Competitive neural network has been widely used in the pattern recognition,classification and other aspects,and shows the great advantages compared with the traditional clustering methods.But the competitive neural network is still inadequate in many aspects,and needs to be further improved.Through the introduction of threshold value learning rules,this paper resolves the issue of getting the training error in the face of such network’s blind spot,on the basis of Kohonen’s Learning Vector Quantization Network(LVQ).Finally this paper programs the realization by Matlab.

Key words: Learning Vector Quantization Network(LVQ), threshold value, blind spot