链状分布的神经网络聚类分析
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
针对以往很多聚类方法不适用于链状分布的样品的情况,运用簇的概念确定新的"重心"点,然后将原始意义上的"距离"改用弧度来表征,再运用神经网络来对样品进行聚类。应用著名的"古斯塔夫森十字"图形进行仿真实验。结果显示:该方法算法简洁,运算迅速,条理清晰,分类准确,适合于处理链状分布的聚类问题。
Many of the methods of the clustering do not apply to the sample which has the chainlike distribution.In order to dissolve this problem,we find a new center of gravity by applying the concept of cluster,and instead of the distance which we always used by radian,then cluster the sample with Neural Network(NN).Finally the famous graph of "The cross of Gustafson" is used as an example.The result shows that this method is compact,rapid,clear and exact.It is fit for dealing with the clustering of the chainlike distribution.
引文
[1]傅京孙.模式识别及其应用[M].北京:科学出版社,1983.
    [2]林柏双.模糊聚类分析在生产管理应用的探讨[J].福建电脑,2003,1:20-21.
    [3]何日光.模糊聚类算法及其应用[J].石油仪器,2004,18[3]:43-46.
    [4]曹辉.地震参数解释的模式识别法[J].石油地球物理勘探,1994,29[2]:140-147.
    [5]王碧泉,陈祖萌.模式识别理论,方法和应用[M].北京:地震出版社,1989.
    [6]Sergios Theodoridis,Konstantinos Koutroumbas.Pattern recognition[M].2nd ed.Beijing:China Machine Press,2003.
    [7]Hathaway R J,Bezdek J C.Local convergence of the fuzzy c-means algorithms[J].Pattern Recognition,1986,19[6]:477-480.
    [8]Hoppner F,Klawonn F,Kruse R,et al.Fuzzy cluster analysis[M].[S.l.]:J Wiley and Sons,1999.
    [9]白莉媛,胡声艳,刘素华.一种基于模拟退火和遗传算法的模糊聚类方法[J].计算机工程与应用,2005,41[9]:56-58.
    [10]刘小芳.点密度加权FCM算法的聚类有效性研究[J].计算机工程与应用,2006,42[15]:20-22.
    [11]袁方,孟增辉,于戈.对k-means聚类算法的改进[J].计算机工程与应用,2004,40[36]:177-178.

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心