Online Community Identification over Heterogeneous Attributed Directed Graphs

The creating of communities has resulted in the astonishing increase in many areas. Especially in the area of social networks, it has wide applications in the domains such as product recommendation, setting up social events, online games etc. The applications are relied on effective solutions for retrieving communities online. In this way, a great deal of research has been conducted on yielding communities. Unfortunately, the state-of-the-art community identity methods which aim to find out communities containing the query nodes, only consider topological structure, but ignore the effect of nodes’ attribute, direction between nodes, and nodes’ information across heterogeneous graphs, lead to communities with poor cohesion. Thus, we address the problem of discovering communities online, across heterogeneous directed attributed graphs. We first propose an online method to match pairs of users in heterogeneous graphs and combine them into a new one. Then we propose IC-ADH, a novel framework of retrieving communities in the new directed attributed graph. Extensive experiments demonstrate the effectiveness of our proposed solution across heterogeneous directed attributed graphs.