A Dimensionality Reduction Framework for Detection of Multiscale Structure in Heterogeneous Networks

被引:9
作者
Shen, Hua-Wei [1 ]
Cheng, Xue-Qi [1 ]
Wang, Yuan-Zhuo [1 ]
Chen, Yixin [2 ]
机构
[1] Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China
[2] Washington Univ, Dept Comp Sci, St Louis, MO 63130 USA
基金
北京市自然科学基金; 中国国家自然科学基金;
关键词
graph clustering; community structure; Laplacian matrix; modularity matrix; dimensionality reduction; COMMUNITY STRUCTURE; COMPLEX NETWORKS; IDENTIFICATION;
D O I
10.1007/s11390-012-1227-y
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Graph clustering has been widely applied in exploring regularities emerging in relational data. Recently, the rapid development of network theory correlates graph clustering with the detection of community structure, a common and important topological characteristic of networks. Most existing methods investigate the community structure at a single topological scale. However, as shown by empirical studies, the community structure of real world networks often exhibits multiple topological descriptions, corresponding to the clustering at different resolutions. Furthermore, the detection of multiscale community structure is heavily affected by the heterogeneous distribution of node degree. It is very challenging to detect multiscale community structure in heterogeneous networks. In this paper, we propose a novel, unified framework for detecting community structure from the perspective of dimensionality reduction. Based on the framework, we first prove that the well-known Laplacian matrix for network partition and the widely-used modularity matrix for community detection are two kinds of covariance matrices used in dimensionality reduction. We then propose a novel method to detect communities at multiple topological scales within our framework. We further show that existing algorithms fail to deal with heterogeneous node degrees. We develop a novel method to handle heterogeneity of networks by introducing a rescaling transformation into the covariance matrices in our framework. Extensive tests on real world and artificial networks demonstrate that the proposed correlation matrices significantly outperform Laplacian and modularity matrices in terms of their ability to identify multiscale community structure in heterogeneous networks.
引用
收藏
页码:341 / 357
页数:17
相关论文
共 55 条
[1]   Modularity-maximizing graph communities via mathematical programming [J].
Agarwal, G. ;
Kempe, D. .
EUROPEAN PHYSICAL JOURNAL B, 2008, 66 (03) :409-418
[2]   Link communities reveal multiscale complexity in networks [J].
Ahn, Yong-Yeol ;
Bagrow, James P. ;
Lehmann, Sune .
NATURE, 2010, 466 (7307) :761-U11
[3]  
[Anonymous], 2008, P 17 INT C WORLD WID, DOI DOI 10.1145/1367497.1367591
[4]   Synchronization reveals topological scales in complex networks [J].
Arenas, A ;
Díaz-Guilera, A ;
Pérez-Vicente, CJ .
PHYSICAL REVIEW LETTERS, 2006, 96 (11)
[5]   Optimal map of the modular structure of complex networks [J].
Arenas, A. ;
Borge-Holthoefer, J. ;
Gomez, S. ;
Zamora-Lopez, G. .
NEW JOURNAL OF PHYSICS, 2010, 12
[6]   Analysis of the structure of complex networks at different resolution levels [J].
Arenas, A. ;
Fernandez, A. ;
Gomez, S. .
NEW JOURNAL OF PHYSICS, 2008, 10
[7]   Evaluating local community methods in networks [J].
Bagrow, James P. .
JOURNAL OF STATISTICAL MECHANICS-THEORY AND EXPERIMENT, 2008,
[8]   Fast unfolding of communities in large networks [J].
Blondel, Vincent D. ;
Guillaume, Jean-Loup ;
Lambiotte, Renaud ;
Lefebvre, Etienne .
JOURNAL OF STATISTICAL MECHANICS-THEORY AND EXPERIMENT, 2008,
[9]   Complex networks: Structure and dynamics [J].
Boccaletti, S. ;
Latora, V. ;
Moreno, Y. ;
Chavez, M. ;
Hwang, D. -U. .
PHYSICS REPORTS-REVIEW SECTION OF PHYSICS LETTERS, 2006, 424 (4-5) :175-308
[10]   On modularity clustering [J].
Brandes, Ulrik ;
Delling, Daniel ;
Gaertler, Marco ;
Goerke, Robert ;
Hoefer, Martin ;
Nikoloski, Zoran ;
Wagner, Dorothea .
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2008, 20 (02) :172-188