期刊文献+
任意字段
题名或关键词
题名
关键词
文摘
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息

学习非唯一的最佳聚类数 被引量：1

Learning a family of intrinsic cluster numbers
原文传递
导出
摘要 确定“最佳聚类数”一直是聚类算法面临的一个难题。为了确定一族合理的聚类数而不是单个聚类数,提出了一种基于谱分析的算法,并能处理较为复杂的数据集。该算法构建了数据点之间的相似度图,在不同的分析粒度下,用图上的“随机游走”来传播相似度,采用了一个新的评判标准,“广义特征差”来寻找聚类数族。实验结果表明该算法在聚类数不唯一的情况下能够有效地确定聚类数,并且和其他几种算法相比具有较优的计算复杂度。 A family of intrinsic cluster numbers, rather than a single cluster number, is determined, using a spectral analysis-based algorithm. The algorithm works not only on simple data sets, but also on more complicated ones. The algorithm constructs an affinity graph which is then modified by.a multi-granularity analysis and a random walk on the graph. A generalized eigengap is defined to determine the cluster number family. Tests show that the algorithm is more effective than previous algorithms and is less complex.
作者 郑欣 林学訚
机构地区 清华大学计算机科学与技术系
出处 《清华大学学报（自然科学版）》 EI CAS CSCD 北大核心 2006年第7期1282-1285,共4页 Journal of Tsinghua University(Science and Technology)
基金 国家"九七三"基础研究基金项目(2002CB312101)
关键词 模式识别 聚类 聚类数自动确定 pattern recognition clustering automatic determination of cluster number
分类号 TP391.4 [自动化与计算机技术—计算机应用技术]
  • 相关文献

参考文献7

  • 1Biernacki C,Govaert G.Choosing Models in Model-based Clustering and Discriminant Analysis[R].Technical Report RR-3509,Rhone-Alpes:INRIA.1999.
  • 2Figueiredo M,Jain A.Unsupervised learning of finite mixture models[J].IEEE Trans on Pattern Analysis and Machine Learning,2002,24(3):381-396.
  • 3Hamerly G,Elkan C.Learning the k in K-means[C]∥ Thrun S,Saul L K,Scholkopf B.Advances in Neural Information Processing Systems 16.Cambridge,MA:MIT Press,2004:281-288.
  • 4Meila M,XU Lei.Multiway Cuts and Spectral Clustering[R].Technical Report No.442,Seattle,WA:U.Washington,2003.
  • 5Ben-Hur A,Elisseeff A,Guyou I.A stability based method for discovering structure in clustered data[C]∥ Altman R B,Dunker A K,Hunter L,et al.Pacific Symposium on Biocomputing.Singapore:World Scientific Press,2002:6-17.
  • 6Lange T,Braun M,Roth V,et al.Stability-based model selection[C]∥ Becker S,Thrun S,Obermayer K.Advances in Neural Information Processing Systems 15.Cambridge,MA:MIT Press,2003:617-624.
  • 7Ng A,Jordan M,Weiss Y.On spectral clustering:analysis and an algorithm[C]∥ Dietterich T G,Becker S,Ghahramani Z.Advances in Neural Information Processing Systems (NIPS) 14.Cambridge,MA:MIT Press,2002:849-856.

同被引文献28

  • 1Wu Z, Leahy R. An optimal graph theoretic approach to data clustering: theory and it's application to image segmentation [J]. IEEE Trans. Pattern Anal,1993,15(11) :1101-1113.
  • 2Kahng H L. A New spectral methods for ratio cut partitioning and clustering [J]. IEEE Transactions on Computed Aided De- sign, 1992,11 (9) :1047-1085.
  • 3Chung F. Spectral graph theory[C]//Conference Board of the Mathematical Sciences. Washington, 1997.
  • 4Shi J, Malik J. Normalized cuts and image segmentation [J].IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 2000,22 (8) : 888-905.
  • 5Kannan R, Vempala S, Vetta A. On clusterings: good, bad and spectral [C] // Proceedings of the Annual IEEE Symposium on Foundations of Computer Science. Redondo, IEEE Press, 2000: 367-377.
  • 6Ng A, Jordan M, Weiss Y. On spectral clustering: analysis and an algorithm[C]//Dietterich T, Becker S, Ghahramani Z, eds. Advances in Neural Information Processing Systems. MIT Press, 2002.
  • 7Ding C, He X,Zha H, et al. A min-max cut algorithm for graph partitioning and data clustering[C]//Proceedings of the first IEEE International Conference on Data Mining(ICDM), 2001. Washington, DC, USA:IEEE Computer Society, 2001 : 107-114.
  • 8Meila M, Shi J. A random walks view of spectral segmentation [C]//Proceedings of AISTATS. Florida, 2001 : 873-879.
  • 9Belkin M. Problems of Learning on Manifolds[D]. University of Chicago, 2003.
  • 10Belkin M,Niyogi P. Laplacian eigenmaps for dimensionality re- duction and data representation[J]. Neural Computation, 2003, 15(6) : 1373-1396.

引证文献1

二级引证文献1

清华大学学报（自然科学版）
2006年 第7期

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部