Sampling Graphical Networks via Conditional Independence Coupling of Markov Chains

被引:0
作者
Li, Guichong [1 ,2 ]
机构
[1] Univ Ottawa, Dept Comp Sci, Ottawa, ON, Canada
[2] NPD Grp, Data & Syst Integrat, Port Washington, NY 11050 USA
来源
ADVANCES IN ARTIFICIAL INTELLIGENCE, AI 2016 | 2016年 / 9673卷
关键词
Sampling online social networks; Markov Chain Monte Carlo; Metropolis-Hastings algorithm; Coupling of Markov Chains; Conditional independence;
D O I
10.1007/978-3-319-34111-8_36
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Markov Chain Monte Carlo (MCMC) methods have been used for sampling Online SNs. The main drawbacks are that traditional MCMC techniques such as the Metropolis-Hastings Random Walk (MHRW) suffer from slow mixing rates, and the resulting sample is usually approximate. An appealing solution is to adapt the MHRW sampler to probability coupling techniques for perfect sampling. While this MHRW coupler is theoretically advanced, it is inapplicable for sampling large SNs in practice. We develop a new coupling algorithm, called Conditional Independence Coupler (CIC), which improves existing coupling techniques by adopting a new coalescence condition, called Conditional Independence (CI), for efficient coalescence detection. The proposed CIC algorithm is outstandingly scalable for sampling large SNs without any bias as compared to previous traditional MCMC sampling algorithms.
引用
收藏
页码:298 / 303
页数:6
相关论文
共 50 条
  • [21] Online Streaming Feature Selection via Conditional Independence
    You, Dianlong
    Wu, Xindong
    Shen, Limin
    He, Yi
    Yuan, Xu
    Chen, Zhen
    Deng, Song
    Ma, Chuan
    APPLIED SCIENCES-BASEL, 2018, 8 (12):
  • [22] Conditional Independence Testing via Latent Representation Learning
    Duong, Bao
    Nguyen, Thin
    2022 IEEE INTERNATIONAL CONFERENCE ON DATA MINING (ICDM), 2022, : 121 - 130
  • [23] A note on the relationship between conditional and unconditional independence and its extensions for Markov kernels
    Nogales, A. G.
    Perez, P.
    STATISTICA NEERLANDICA, 2019, 73 (03) : 320 - 332
  • [24] Functional compatibility, Markov chains and Gibbs sampling with improper posteriors
    Hobert, J
    Casella, G
    DIMENSION REDUCTION, COMPUTATIONAL COMPLEXITY AND INFORMATION, 1998, 30 : 29 - 29
  • [25] Kernel partial correlation: a novel approach to capturing conditional independence in graphical models for noisy data
    Oh, Jihwan
    Zheng, Faye
    Doerge, R. W.
    Chun, Hyonho
    JOURNAL OF APPLIED STATISTICS, 2018, 45 (14) : 2677 - 2696
  • [26] The sampling properties of conditional independence graphs for I(1) structural VAR models
    Wilson, Granville Tunnicliffe
    Reale, Marco
    JOURNAL OF TIME SERIES ANALYSIS, 2008, 29 (05) : 802 - 810
  • [27] Product disintegrations: A law of large numbers via conditional independence
    Borsato, Luisa
    Horta, Eduardo
    Souza, Rafael Rigao
    STATISTICS & PROBABILITY LETTERS, 2024, 208
  • [28] Testing conditional independence via integrating-up transform
    Liu, Yi
    Wang, Qihua
    Liu, Xiaohui
    STATISTICS, 2018, 52 (04) : 734 - 749
  • [29] ESTIMATING STANDARD ERRORS FOR IMPORTANCE SAMPLING ESTIMATORS WITH MULTIPLE MARKOV CHAINS
    Roy, Vivekananda
    Tan, Aixin
    Flegal, James M.
    STATISTICA SINICA, 2018, 28 (02) : 1079 - 1101
  • [30] Some characterizations of minimal Markov basis for sampling from discrete conditional distributions
    Akimichi Takemura
    Satoshi Aoki
    Annals of the Institute of Statistical Mathematics, 2004, 56 : 1 - 17