PSEUDO-LIKELIHOOD METHODS FOR COMMUNITY DETECTION IN LARGE SPARSE NETWORKS

被引:242
作者
Amini, Arash A. [1 ]
Chen, Aiyou [2 ]
Bickel, Peter J. [3 ]
Levina, Elizaveta [1 ]
机构
[1] Univ Michigan, Dept Stat, Ann Arbor, MI 48109 USA
[2] Google Inc, Mountain View, CA 94043 USA
[3] Univ Calif Berkeley, Dept Stat, Berkeley, CA 94760 USA
基金
美国国家科学基金会;
关键词
Community detection; network; pseudo-likelihood; STOCHASTIC BLOCKMODELS; DISTRIBUTIONS; CONSISTENCY; PREDICTION; SUCCESSES; NUMBER; GRAPHS; MODELS;
D O I
10.1214/13-AOS1138
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Many algorithms have been proposed for fitting network models with communities, but most of them do not scale well to large networks, and often fail on sparse networks. Here we propose a new fast pseudo-likelihood method for fitting the stochastic block model for networks, as well as a variant that allows for an arbitrary degree distribution by conditioning on degrees. We show that the algorithms perform well under a range of settings, including on very sparse networks, and illustrate on the example of a network of political blogs. We also propose spectral clustering with perturbations, a method of independent interest, which works well on sparse networks where regular spectral clustering fails, and use it to provide an initial value for pseudo-likelihood. We prove that pseudo-likelihood provides consistent estimates of the communities under a mild condition on the starting value, for the case of a block model with two communities.
引用
收藏
页码:2097 / 2122
页数:26
相关论文
共 38 条
  • [1] Airoldi EM, 2008, J MACH LEARN RES, V9, P1981
  • [2] [Anonymous], JMLR WORKSHOP C P
  • [3] [Anonymous], 2005, P WWW 2005 WORKSH WE
  • [4] [Anonymous], ASYMPTOTIC NORMALITY
  • [5] [Anonymous], ADV NEURAL INFORM PR
  • [6] [Anonymous], PSEUDOLIKELIHOOD M S
  • [7] Efficient and principled method for detecting communities in networks
    Ball, Brian
    Karrer, Brian
    Newman, M. E. J.
    [J]. PHYSICAL REVIEW E, 2011, 84 (03)
  • [8] BESAG J, 1974, J ROY STAT SOC B MET, V36, P192
  • [9] Bickel P. J., 2007, Mathematical statistics, basic ideas and selected topics
  • [10] THE METHOD OF MOMENTS AND DEGREE DISTRIBUTIONS FOR NETWORK MODELS
    Bickel, Peter J.
    Chen, Aiyou
    Levina, Elizaveta
    [J]. ANNALS OF STATISTICS, 2011, 39 (05) : 2280 - 2301