Detecting overlapping communities with triangle-based rough local expansion method

Abstract

Overlapping communities structures could effectively reveal the internal relationships in real networks, especially on the ownership problems on the nodes in overlapping areas between communities. Hence, the overlapping community detection research becomes a hotspot topic on graph mining in the decade, while the local expansion methods based on structural fitness function could simultaneously discover overlapped and hierarchical structures. Aimed at community drift and redundant calculation problems on general local expansion methods, the paper presents a novel local expansion method based on rough neighborhood that carries out the heuristic technology by the community seed inspiration and the triangle optimization on the boundary domain. The method could directly generate natural overlaps between communities, reduce the computational complexity and improve the detection on the overlapped boundary area. Finally, the experimental results on some real networks also show that the rough expansion method based on triangle optimization could be more effective in detecting the overlapping structures.