ESTIMATING AND UNDERSTANDING EXPONENTIAL RANDOM GRAPH MODELS

被引:191
作者
Chatterjee, Sourav [1 ]
Diaconis, Persi [1 ]
机构
[1] Stanford Univ, Dept Stat, Stanford, CA 94305 USA
基金
美国国家科学基金会;
关键词
Random graph; Erdos-Renyi; graph limit; exponential random graph models; parameter estimation; CONVERGENT SEQUENCES; MATRICES; RANK;
D O I
10.1214/13-AOS1155
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
We introduce a method for the theoretical analysis of exponential random graph models. The method is based on a large-deviations approximation to the normalizing constant shown to be consistent using theory developed by Chatterjee and Varadhan [European J. Combin. 32 (2011) 1000-1017]. The theory explains a host of difficulties encountered by applied workers: many distinct models have essentially the same MLE, rendering the problems "practically" ill-posed. We give the first rigorous proofs of "degeneracy" observed in these models. Here, almost all graphs have essentially no edges or are essentially complete. We supplement recent work of Bhamidi, Bresler and Sly [2008 IEEE 49th Annual IEEE Symposium on Foundations of Computer Science (FOCS) (2008) 803-812 IEEE] showing that for many models, the extra sufficient statistics are useless: most realizations look like the results of a simple Erdos-Renyi model. We also find classes of models where the limiting graphs differ from Erdos-Renyi graphs. A limitation of our approach, inherited from the limitation of graph limit theory, is that it works only for dense graphs.
引用
收藏
页码:2428 / 2461
页数:34
相关论文
共 58 条
[1]   REPRESENTATIONS FOR PARTIALLY EXCHANGEABLE ARRAYS OF RANDOM-VARIABLES [J].
ALDOUS, DJ .
JOURNAL OF MULTIVARIATE ANALYSIS, 1981, 11 (04) :581-598
[2]  
[Anonymous], SOCIAL NETWORK ANAL
[3]  
[Anonymous], 2002, J SOC STRUCTURE
[4]  
[Anonymous], 2003, Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics
[5]  
ARISTOFF D., 2011, PREPRINT
[6]   Testability and Repair of Hereditary Hypergraph Properties [J].
Austin, Tim ;
Tao, Terence .
RANDOM STRUCTURES & ALGORITHMS, 2010, 36 (04) :373-463
[7]   On exchangeable random variables and the statistics of large graphs and hypergraphs [J].
Austin, Tim .
PROBABILITY SURVEYS, 2008, 5 :80-145
[8]   STATISTICAL-ANALYSIS OF NON-LATTICE DATA [J].
BESAG, J .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES D-THE STATISTICIAN, 1975, 24 (03) :179-195
[9]   Mixing time of exponential random graphs [J].
Bhamidi, Shankar ;
Bresler, Guy ;
Sly, Allan .
PROCEEDINGS OF THE 49TH ANNUAL IEEE SYMPOSIUM ON FOUNDATIONS OF COMPUTER SCIENCE, 2008, :803-+
[10]  
Bollobas B., 2001, RANDOM GRAPHS, DOI 10.1017/CBO9780511814068