SAMPLING AND ESTIMATION FOR (SPARSE) EXCHANGEABLE GRAPHS

被引:20
|
作者
Veitch, Victor [1 ]
Roy, Daniel M. [2 ]
机构
[1] Columbia Univ, Dept Stat, 1255 Amsterdam Ave, New York, NY 10027 USA
[2] Sidney Smith Hall, Dept Stat Sci, 100 St George St, Toronto, ON M5S 3G3, Canada
关键词
Network analysis; sampling; nonparametric estimation; L-P THEORY; CONVERGENCE; MODELS;
D O I
10.1214/18-AOS1778
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
Sparse exchangeable graphs on R+, and the associated graphex framework for sparse graphs, generalize exchangeable graphs on N, and the associated graphon framework for dense graphs. We develop the graphex framework as a tool for statistical network analysis by identifying the sampling scheme that is naturally associated with the models of the framework, formalizing two natural notions of consistent estimation of the parameter (the graphex) underlying these models, and identifying general consistent estimators in each case. The sampling scheme is a modification of independent vertex sampling that throws away vertices that are isolated in the sampled subgraph. The estimators are variants of the empirical graphon estimator, which is known to be a consistent estimator for the distribution of dense exchangeable graphs; both can be understood as graph analogues to the empirical distribution in the i.i.d. sequence setting. Our results may be viewed as a generalization of consistent estimation via the empirical graphon from the dense graph regime to also include sparse graphs.
引用
收藏
页码:3274 / 3299
页数:26
相关论文
共 50 条
  • [31] Estimation of covariance matrix via the sparse Cholesky factor with lasso
    Chang, Changgee
    Tsay, Ruey S.
    JOURNAL OF STATISTICAL PLANNING AND INFERENCE, 2010, 140 (12) : 3858 - 3873
  • [32] The dual channel sinewave model: Co-prime sparse sampling, parameter estimation, and the Cramer-Rao Bound
    Negusse, Senay
    Handel, Peter
    MEASUREMENT, 2012, 45 (09) : 2254 - 2263
  • [33] SPARSE REGULAR RANDOM GRAPHS: SPECTRAL DENSITY AND EIGENVECTORS
    Dumitriu, Ioana
    Pal, Soumik
    ANNALS OF PROBABILITY, 2012, 40 (05) : 2197 - 2235
  • [34] Thinned random measures for sparse graphs with overlapping communities
    Ricci, Federica Zoe
    Guindani, Michele
    Sudderth, Erik B.
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
  • [35] DEGREE AND CLUSTERING COEFFICIENT IN SPARSE RANDOM INTERSECTION GRAPHS
    Bloznelis, Mindaugas
    ANNALS OF APPLIED PROBABILITY, 2013, 23 (03) : 1254 - 1289
  • [36] Convergent Sequences of Sparse Graphs: A Large Deviations Approach
    Borgs, Christian
    Chayes, Jennifer
    Gamarnik, David
    RANDOM STRUCTURES & ALGORITHMS, 2017, 51 (01) : 52 - 89
  • [37] Generalized designs on graphs: Sampling, spectra, symmetries
    Steinerberger, Stefan
    JOURNAL OF GRAPH THEORY, 2020, 93 (02) : 253 - 267
  • [38] Exact sampling of graphs with prescribed degree correlations
    Bassler, Kevin E.
    Del Genio, Charo I.
    Erdos, Peter L.
    Miklos, Istvan
    Toroczkai, Zoltan
    NEW JOURNAL OF PHYSICS, 2015, 17
  • [39] SAMPLING THEORY FOR GRAPH SIGNALS ON PRODUCT GRAPHS
    Varma, Rohan A.
    Kovacevic, Jelena
    2018 IEEE GLOBAL CONFERENCE ON SIGNAL AND INFORMATION PROCESSING (GLOBALSIP 2018), 2018, : 768 - 772
  • [40] Computation of Graphlet Orbits for Nodes and Edges in Sparse Graphs
    Hocevar, Tomaz
    Demsar, Janez
    JOURNAL OF STATISTICAL SOFTWARE, 2016, 71 (10): : 1 - 24