User Topic Preferences Based Influence Maximization in Overlapped Networks

With the rapid development of online social networks, there is a trend of information interconnection among online social networks. User's social behaviors are no longer confined to a single network but interact in the form of overlapping networks. Studying the problem of maximizing the influence of overlapping networks can not only make the influence spread to the extent that a single network cannot reach, but also effectively control the cost of advertising while maintaining the scope of diffusion. However, the existing models of user identification and influence propagation in overlapping networks are still not efficient enough. For this reason, this paper studies the problem of maximizing influence in overlapping networks based on user interests. First, network coupling is carried out through the “bridge” role of overlapping network users; secondly, independent cascade model is used. On the basis of this, a User Interest-based Influence Propagation Model of the Overlay Network (UI-IPM) is designed. Finally, based on the UI-IPM model, a heuristic algorithm combined with the greedy algorithm is designed to maximize the impact of overlapping networks (UI-IPM) and achieve Influence Maximization of the Overlay Network (IMON) mining seed nodes. The experimental results show the effectiveness of IMON algorithm in terms of influence sphere and time efficiency, and also verify the efficiency of seed nodes mining in overlapping network environment compared with a single network environment.