BI-COMDET: Community Detection in Bipartite Networks

Extracting hidden communities from bipartite networks witnessed a determined effort. In this respect, different streams of research relied on bipartite networks to unveil communities. In this paper, we introduce a new approach, called BI-COMDET, that aims to an efficient community detection in bipartite networks. The main trust of the introduced approach is that it stresses on the importance of grouping two types of nodes in communities having a full connection between its nodes. The quality of the unveiled communities, is assessed through some metrics borrowed from the FCA community, to wit modularity, overlapping and stability. These metrics are then aggregated through the use of multi-criteria method to elect the most pertinent bi-comunity from some candidates. Carried out experiments show that BI-COMDET sharply outperforms its competitors in terms of modularity, Conductance and intra/inter density. (C) 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of KES International.