首页 | 本学科首页   官方微博 | 高级检索  
     

基于模糊K-调和均值的单词-文档谱聚类方法
引用本文:刘娜,肖智博,鲁明羽.基于模糊K-调和均值的单词-文档谱聚类方法[J].控制与决策,2012,27(4):501-506.
作者姓名:刘娜  肖智博  鲁明羽
作者单位:大连海事大学信息科学技术学院;大连工业大学信息科学与工程学院
基金项目:国家自然科学基金项目(61175053,61073133,60973067);教育部创新团队及重点科研培育项目(2011ZD010)
摘    要:在分析单词-文档谱聚类方法的基本步骤,找出其对初始值敏感的根本原因的基础上,提出一种基于模糊-调和均值的单词-文档谱聚类方法.首先从矩阵相似的角度对谱聚类中的Laplacian矩阵进行处理,使其满足对初始值不敏感的条件;然后通过加入模糊的概念,用模糊K-调和均值算法代替K-均值算法,使聚类结果对初始值不敏感.实验结果表明,所提出的方法不仅使聚类结果对初始值不敏感,而且在一定程度上提高了数据的鲁棒性.

关 键 词:谱聚类  K-均值  K-调和均值  模糊-调和均值
收稿时间:2010/11/10 0:00:00
修稿时间:2011/1/14 0:00:00

Spectral co-clustering documents and words based on fuzzy K-Harmonic means
LIU Na XIAO Zhi-bo LU Ming-yu.Spectral co-clustering documents and words based on fuzzy K-Harmonic means[J].Control and Decision,2012,27(4):501-506.
Authors:LIU Na XIAO Zhi-bo LU Ming-yu
Affiliation:1(1.College of Information Science & Technology,Dalian Maritime University,Dalian 116026,China;2.College of Information Science & Engineering,Dalian Polytechnic University,Dalian 116034,China.)
Abstract:Based on analysising the main step of spectral clustering and finding out its cause of sensitive to the initialization,a method of spectral co-clustering documents and words based on fuzzy K-harmonic means is proposed.Firstly,the matrix which is insensitive to the initialization is constructed.Then fuzzy K-harmonic means algorithm is used instead of K-means algorithm.The experiment result shows that the proposed method not only is initialization insensitive,but also can improve the accuracy and robustness of clustering results.
Keywords:spectral clustering  K-means  K-harmonic means  fuzzy-harmonic means
本文献已被 CNKI 等数据库收录!
点击此处可从《控制与决策》浏览原始摘要信息
点击此处可从《控制与决策》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号