三种典型的基于图分割的谱聚类方法比较

被引:2
作者
王娜 [1 ]
杜海峰 [2 ,1 ]
庄健 [1 ]
余进涛 [1 ]
王孙安 [1 ]
机构
[1] 西安交通大学机械学院,现代设计和转子轴承系统教育部重点试验室
[2] 西安交通大学公共政策与管理学院
关键词
聚类; 图分割; 谱聚类; 谱图理论;
D O I
10.16182/j.cnki.joss.2009.11.071
中图分类号
TP181 [自动推理、机器学习];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
在分析谱聚类实现思路和已有算法基础上,对规范切判据,最小最大切判据和自动确定聚类数目的谱聚类典型算法进行了研究和应用,通过理论分析算法各自实现机理的联系与区别,讨论它们各自的聚类特点,并利用UCI(University of California,Irvine)机器学习数据集试验对比了三种算法的聚类效果。发现谱聚类算法实现数据聚类的有效性,以及参数及相似度度量对算法性能有很大影响,在此基础上提出了算法用于解决可建模为模式识别的工程问题的可行思路,为工程实践提供了借鉴。
引用
收藏
页码:3316 / 3320
页数:5
相关论文
共 11 条
[1]  
UCI repositoryof machine learning databases. BLAKE C L,KEOGH E,MERZ C J. . 1998
[2]  
‘Matrix Computations'. Golub, G. H.,and C. F. Vanloan. . 1983
[3]  
Normalized cuts and image segmentation. J Shi,J Malik. IEEE Transactions on Pattern Analysis and Machine Intelligence . 2000
[4]  
A comparison of spectral clustering algorithms. VERMA D,MEILA M. . 2003
[5]  
A New Spectral Algorithm for Large Training Sets. PRIETO R,JIANG J,CHOI C H. Proceedmgs of the Second International Conference on Machine Learning and Cybernetics . 2003
[6]  
An optimal graph theoretic approach to data clustering: theory and its application to image segmentation. Wu,Z.,Leahy,R. IEEE Transactions on Pattern Analysis and Machine Intelligence . 1993
[7]  
Some methods for classification and analysis of multivariate observations. MacQueen J. Proceedings of the Fifth Berkeley Symposium on Mathematics, Statistics and Science . 1967
[8]  
Automatic Determination of the Number of Clusters Using Spectral Algorithms. Sanguinetti G,Laidler J,Lawrence N. Proc of IEEE Machine Learning for Signal Processing 2005 . 2005
[9]  
On spectral clustering: Analysis and an algorithm. Ng A.Y,Jordan M. I,and Weiss Y. Advances in Neural Information Processing Systems . 2002
[10]  
A rain-max cut for graph partitioning and data clustering. DING C,REN X F,ZHA H,et al. IEEE Int.Conf. on Data Mining Advanced Materials and Processing . 2001