Bayesian consensus clustering in multiplex networks

Abstract

Multiplex networks are immanently characterized with heterogeneous relations among vertices. In this paper, we develop Bayesian consensus stochastic block modeling for multiplex networks. The posterior distribution of the model is approximated via Markov chain Monte Carlo, and a Gibbs sampler is derived in detail. The model allows both integrated analysis of heterogeneous relations, thus providing more accurate block assignments, and simultaneously handling uncertainty in the model parameters. Motivated by the fact that the symmetry in physics plays a crucial role, we discuss also the symmetry in statistics, which is nowadays commonly known as exchangeability—the concept that has recently transformed the field of statistical network analysis.

Document Details

Document Type
Pub Defense Publication
Publication Date
Oct 01, 2019
Source ID
10.1063/1.5120503

Entities

People

  • Ljupco Kocarev
  • Petar Jovanovski

Organizations

  • Macedonian Academy of Sciences and Arts
  • Office of Naval Research
  • Office of Naval Research Global
  • Ss. Cyril and Methodius University of Skopje
  • University of California, San Diego

Tags

Fields of Study

  • Mathematics

Readers

  • Distributed Systems and Data Platform Development
  • Graph Algorithms and Convex Optimization.
  • Statistical inference.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference