Exploring influence maximization in online and offline double-layer propagation scheme

Information propagation in network environment is a widely studied research topic, especially in Online Social Networks (OSNs), where the problem has gained significant popularity. Recent studies attempt to pick up the key nodes, who could maximize the network influence in OSNs. However, in addition to propagation in OSNs, another information propagation way is through words of mouth among people in the offline mobile network, which is an indispensable factor and is not considered in most cases. Hence, the information propagation in both online social network and offline mobile network is a new valuable scheme. In this paper, we propose an Information Maximization strategy in Online and Offline double-layer Propagation scheme (IMOOP), where we first form the topological graph for online social network and offline connection graph of probability, respectively. Then, the two layers are compressed into a single-layer communication graph. We further prove that the influence maximization in double-layer propagation scheme is NP-hard, then we describe practical greedy heuristics for the resulting NP-hard problems and compute their approximation ratios. Our experiments with real mobility datasets (Brightkite, Gowalla and Foursquare) show that, the proposed propagation scheme achieves a higher information cover ratio, compared with the other propagation methods.