
For FullText PDF, please login, if you are a member of IEICE,
or go to Pay Per View on menu list, if you are a nonmember of IEICE.

Extreme Maximum Margin Clustering
Chen ZHANG ShiXiong XIA Bing LIU Lei ZHANG
Publication
IEICE TRANSACTIONS on Information and Systems
Vol.E96D
No.8
pp.17451753 Publication Date: 2013/08/01 Online ISSN: 17451361
DOI: 10.1587/transinf.E96.D.1745 Print ISSN: 09168532 Type of Manuscript: PAPER Category: Artificial Intelligence, Data Mining Keyword: maximum margin clustering, unsupervised learning, extreme learning machine (ELM), random feature mapping,
Full Text: PDF>>
Summary:
Maximum margin clustering (MMC) is a newly proposed clustering method that extends the largemargin computation of support vector machine (SVM) to unsupervised learning. Traditionally, MMC is formulated as a nonconvex integer programming problem which makes it difficult to solve. Several methods rely on reformulating and relaxing the nonconvex optimization problem as semidefinite programming (SDP) or secondorder cone program (SOCP), which are computationally expensive and have difficulty handling largescale data sets. In linear cases, by making use of the constrained concaveconvex procedure (CCCP) and cutting plane algorithm, several MMC methods take linear time to converge to a local optimum, but in nonlinear cases, time complexity is still high. Since extreme learning machine (ELM) has achieved similar generalization performance at much faster learning speed than traditional SVM and LSSVM, we propose an extreme maximum margin clustering (EMMC) algorithm based on ELM. It can perform well in nonlinear cases. Moreover, the kernel parameters of EMMC need not be tuned by means of random feature mappings. Experimental results on several realworld data sets show that EMMC performs better than traditional MMC methods, especially in handling largescale data sets.

