作者: Yihong Wu , Jiaming Xu , Bruce Hajek
DOI: 10.1109/ACSSC.2015.7421303
关键词: Cluster (physics) 、 Maximum likelihood 、 Stochastic block model 、 Relaxation (approximation) 、 Applied mathematics 、 Conjecture 、 Combinatorics 、 Binary number 、 Semidefinite programming 、 Mathematics
摘要: Resolving a conjecture of Abbe, Bandeira and Hall, the authors have recently shown that semidefinite programming (SDP) relaxation maximum likelihood estimator achieves sharp threshold for exactly recovering community structure under binary stochastic block model two equal-sized clusters. Extending proof techniques, in this paper it is SDP relaxations also achieve recovery with fixed number