Exchangeable Random Measures for Sparse and Modular Graphs with Overlapping Communities

Abstract : We propose a novel statistical model for sparse networks with overlapping community structure. The model is based on representing the graph as an exchangeable point process, and naturally generalizes existing probabilistic models with overlapping block-structure to the sparse regime. Our construction builds on vectors of completely random measures, and has interpretable parameters, each node being assigned a vector representing its level of affiliation to some latent communities. We develop methods for simulating this class of random graphs, as well as to perform posterior inference. We show that the proposed approach can recover interpretable structure from two real-world networks and can handle graphs with thousands of nodes and tens of thousands of edges.
Liste complète des métadonnées

Cited literature [73 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01270854
Contributor : Adrien Todeschini <>
Submitted on : Monday, February 8, 2016 - 3:35:30 PM
Last modification on : Tuesday, April 17, 2018 - 9:08:10 AM
Document(s) archivé(s) le : Saturday, November 12, 2016 - 1:28:56 PM

File

1602.02114v1.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01270854, version 1
  • ARXIV : 1602.02114

Collections

Citation

Adrien Todeschini, Francois Caron. Exchangeable Random Measures for Sparse and Modular Graphs with Overlapping Communities. 2016. 〈hal-01270854〉

Share

Metrics

Record views

307

Files downloads

230