首页 | 本学科首页   官方微博 | 高级检索  
     

聚类趋势问题的研究综述*
引用本文:褚娜,马利庄,王彦.聚类趋势问题的研究综述*[J].计算机应用研究,2009,26(3):801-803.
作者姓名:褚娜  马利庄  王彦
作者单位:1. 上海交通大学,电子信息与电气工程学院,计算机科学与工程系,上海,200240
2. 上海交通大学,电子信息与电气工程学院,计算机科学与工程系,上海,200240;上海中医药大学,信息科学与技术中心,上海,201203
基金项目:国家“973”计划资助项目(2006CB504801); 上海市科委世博科技专项资助项目(06dz05815)
摘    要:聚类算法的性能与数据集的结构是密切相关的,虽然目前已经研究出了很多聚类算法,但没有普遍适用的万能聚类算法,欠缺对数据集结构的有效解释。对聚类分析过程中重要的关键性问题,即聚类趋势问题进行了系统性的研究,从统计检验、可视化分析等角度给予了讨论,为数据集的无监督聚类分析提供了合理和有效的前期分析工具。

关 键 词:聚类趋势    聚类分析    统计检验    可视化评估

Research for clustering tendency
CHU N,MA Li zhuang,WANG Yan.Research for clustering tendency[J].Application Research of Computers,2009,26(3):801-803.
Authors:CHU N  MA Li zhuang  WANG Yan
Affiliation:(1.Dept. of Computer Science & Engineering, School of Electronic Information & Electrical Engineering, Shanghai Jiaotong University, Shanghai 200240, China; 2.Center of Information Science & Technology, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China)
Abstract:It is closely related between the performance of clustering algorithms and the structure of data sets. No methods were good enough for all types of data, nor were all methods equally applicable to all problems, and were short of reasonable interpretation to data structure. And systematic researched the clustering tendency, which was one of key problem about clustering analysis. This paper discussed it based on statistic tests, visual analysis and so on, and proved that it could present reasonable and effective analysis tools for unsupervised clustering analysis of data sets.
Keywords:clustering tendency  clustering analysis  statistic tests  visual assessment
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《计算机应用研究》浏览原始摘要信息
点击此处可从《计算机应用研究》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