The Community for Technology Leaders
Green Image
Issue No. 06 - June (2012 vol. 24)
ISSN: 1041-4347
pp: 1025-1035
Zhengguang Chen , Zhejiang Provincial Key Lab. of Service Robot, Zhejiang Univ., Hangzhou, China
Deng Cai , State Key Lab. of CAD&CG, Zhejiang Univ., Hangzhou, China
Jiawei Han , Dept. of Comput. Sci., Univ. of Illinois at Urbana-Champaign, Urbana, IL, USA
Jiajun Bu , Zhejiang Provincial Key Lab. of Service Robot, Zhejiang Univ., Hangzhou, China
Chun Chen , Zhejiang Provincial Key Lab. of Service Robot, Zhejiang Univ., Hangzhou, China
Lijun Zhang , Zhejiang Provincial Key Lab. of Service Robot, Zhejiang Univ., Hangzhou, China
ABSTRACT
Different from traditional one-sided clustering techniques, coclustering makes use of the duality between samples and features to partition them simultaneously. Most of the existing co-clustering algorithms focus on modeling the relationship between samples and features, whereas the intersample and interfeature relationships are ignored. In this paper, we propose a novel coclustering algorithm named Locally Discriminative Coclustering (LDCC) to explore the relationship between samples and features as well as the intersample and interfeature relationships. Specifically, the sample-feature relationship is modeled by a bipartite graph between samples and features. And we apply local linear regression to discovering the intrinsic discriminative structures of both sample space and feature space. For each local patch in the sample and feature spaces, a local linear function is estimated to predict the labels of the points in this patch. The intersample and interfeature relationships are thus captured by minimizing the fitting errors of all the local linear functions. In this way, LDCC groups strongly associated samples and features together, while respecting the local structures of both sample and feature spaces. Our experimental results on several benchmark data sets have demonstrated the effectiveness of the proposed method.
INDEX TERMS
regression analysis, graph theory, pattern clustering, LDCC groups, locally discriminative coclustering, one-sided clustering techniques, interfeature relationships, intersample relationships, sample-feature relationship, bipartite graph, local linear regression, intrinsic discriminative structures, sample space, feature space, local patch, local linear function, fitting errors, Clustering algorithms, Bipartite graph, Matrix decomposition, Partitioning algorithms, Linear regression, Silicon, Mathematical model, local linear regression., Coclustering, clustering, bipartite graph
CITATION
Zhengguang Chen, Deng Cai, Jiawei Han, Jiajun Bu, Chun Chen, Lijun Zhang, "Locally Discriminative Coclustering", IEEE Transactions on Knowledge & Data Engineering, vol. 24, no. , pp. 1025-1035, June 2012, doi:10.1109/TKDE.2011.71
93 ms
(Ver 3.1 (10032016))