[1]WANG Zhong,LIU Gui-quan,CHEN En-hong.A spectral clustering algorithm based on fuzzy Kharmonic means[J].CAAI Transactions on Intelligent Systems,2009,4(2):95-99.
Copy

A spectral clustering algorithm based on fuzzy Kharmonic means

CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume: 4 Number of periods: 2009 2 Page number: 95-99 Column: 学术论文—机器学习 Public date: 2009-04-25
Title:
A spectral clustering algorithm based on fuzzy Kharmonic means
Author(s):
WANG Zhong1; 2;  LIU Gui-quan1; 2; CHEN En-hong1; 2
1.School of Computer Science, University of Science and Technology of China, Hefei 230027,China；
2. Key Laboratory of Software in Computing and Communication, Hefei 230027, China
Keywords:
spectral clustering; fuzzy Kharmonic means; initialization sensitivity; cluster centers
CLC:
TP311
DOI:
-
Abstract:
Spectral clustering is an effective method that is widely used in machine learning. After analyzing the essence of initialization sensitivity in spectral clustering, the fuzzy Kharmonic means (FKHM) algorithm was considered to conquer spectral clustering’s shortcomings, then an spectral clustering algorithm based on FKHM was developed. Compared with the traditional spectral algorithm and the fuzzy cmeans (FCM) algorithm, the suggested algorithm is more sensitive to initial values. The suggested algorithm can not only identify challenging artificial data, but also find stable cluster centers and clustering results, considerably improving clustering precision. Experiments showed that it is an effective and feasible way to improve the performance of spectral clustering algorithms.
References:
［1］FIEDLER M. Algebraic connectivity of graphs［M］. Praha: Czechoslovak Mathematical Journal, 1973:298305.［2］VERMA D, MEILA M. A comparison of spectral clustering algorithms［R］. University of Washington, 2003.
 ［3］FISCHER I, POLAND J. Amplifying the block matrix structure for spectral clustering［C］//Proceedings of the 14th Annual Machine Conference of Belgium and the Netherlands. Manno, Switzerland, 2005:2128.
［4］FOWLKES C, BELONGIE S, CHUNG F, et al. Spectral grouping using the Nystr〖AKo¨〗m method［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, 26(2):217225.
［5］EKIN A, PANKANTI S, HAMPAPUR A. Initializationindependent spectral clustering with applications to automatic video analysis［C］//Proc of IEEE ICASSP. Montreal, Canada,2004:641644.
［6］SHI J B,MALIK J. Normalized cuts and image segmentation［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，2000,22(8):888905.
?［7］NG A Y,JORDAN M L, WEISS Y. On spectral clustering:analysis and an algorithm［C］//Advances in Neural Information Processing Systems. Cambridge: MIT Press,2002:897856.
［8］赵 恒,杨万海,张高煜.模糊KHarmonic Means聚类算法［J］.西安电子科技大学学报,2005,32(4):603606.
ZHAO Heng, YANG Wanhai, ZHANG Gaoyu. Fuzzy Kharmonic means clustering algorithm［J］.Journal of XiDian University,2005,32(4):603606
 ［9］ZHANG B,HSU M,DAYAL U. Kharmonic means—a data clustering algorithm［EB/OL］.［20060112］.http://hpc.isti.cnr.it/~palmeri/datam/articles/HPL1999124.pdf.
［10］HANDL J,KNOWLES J. An evolutionary approach to multiobjective clustering［J］. IEEE Transactions on Evolutionary Computation,2007,11(1):5676.
Similar References:

Memo

-

Last Update: 2009-05-04

Copyright © CAAI Transactions on Intelligent Systems