Identifying the social-balanced densest subgraph from signed social networks

Abstract

Identifying the dense subgraphs from large graphs is important and useful to various social media mining applications. Most of existing works focus on the densest subgraph problem in the unweighted and undirected represented social network which can maximize the average degree over all possible subgraphs. However, considering the frequent signed relationships occurred in real-life social network, this paper introduces the social-balanced densest subgraph problem in signed social network by incorporating the social balance theory. We obtain a novel problem formulation that is to identify the subset of vertices that can maximize the social-balanced density in signed social networks. Further, we propose an efficient approach for identifying the social-balanced densest subgraph based on formal concept analysis. The case study illustrates that our algorithm can efficiently identify the social-balanced densest subgraph for satisfying the specific application's requirements.