Mathematical programming approaches to community detection problems.

  1. F. Temprano Garcia
  2. J. Puerto Albandoz
  3. A.M. Rodríguez-Chía
  4. S. Benati
Actes de conférence:
XXXIX CONGRESO NACIONAL DE ESTADÍSTICA E INVESTIGACIÓN OPERATIVA XIII JORNADAS DE ESTADÍSTICA PÚBLICA

Éditorial: Departamento de Estadística e Investigación Operativa. Universidad de Granada

ISBN: 978-84-09-41628-8

Année de publication: 2022

Pages: 117

Type: Communication dans un congrès

GOOGLE SCHOLAR lock_openAccès ouvert editor

Résumé

We present a general methodology using mathematical optimization to identify overlapping communities in complex networks by maximizing aspiration criteria based on extensions of the Newman and Girvan modularity function. We provide mathematical programming formulations both for a new proposed modularity function and for the extended modularity function by Zhang et al. (2007) to find optimal overlapping communities for small to medium size networks. We also develop a heuristic algorithm for the new proposed modularity function valid for large networks. Experimental results indicate, on the one hand, that optimizing the extended modularity function proposed by Zhang et al. (2007) may produce senseless communities and on the other hand, our results show that the new modularity function is efficient at detecting good clusteringswith overlapping.