RODIN. Repositorio de Objetos de Docencia e Investigación de la Universidad de Cádiz
Publicación
Artículo científico (JournalArticle)
A mathematical programming approach to overlapping community detection
We propose a new optimization model to detect overlapping communities in networks. The model elaborates suggestions contained in Zhang et al. (2007), in which overlapping communities were identified through the use of a fuzzy membership function, calculated as the outcome of a mathematical programming problem. In our approach, we retain the idea of using both mathematical programming and fuzzy membership to detect overlapping communities, but we replace the fuzzy objective function proposed there with another one, based on the Newman and Girvan's definition of modularity. Next, we formulate a new mixed-integer linear programming model to calculate optimal overlapping communities. After some computational tests, we provide some evidence that our new proposal can fix some biases of the previous model, that is, its tendency of calculating communities composed of almost all nodes. Conversely, our new model can reveal other structural properties, such as nodes or communities acting as bridges between communities. Finally, as mathematical programming can be used only for moderate size networks due to its computation time, we proposed two heuristic algorithms to solve the largest instances, that compare favourably to other methodologies. (c) 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Proyecto:
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020 PID2020-114594GB-C22
,
Junta de Andalucía / Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020 P18-FR-1422
,
Junta de Andalucía FEDER-US-1256951
,
AEI PID2020-114594GB-C21
,
Junta de Andalucía FQM-331