Group Influence Maximization in Social Networks

In emerging applications of social networks, groups play a vital role as most decisions are made by groups according to the opinion of the majority therein. This brings the problem of Group Influence Maximization (GIM) which aims to select k initial active nodes for maximizing the expected number of influenced groups. In the paper, we study GIM and focus on activating groups rather than individuals. Observing the known NP-hardness of GIM and the \(\#P\)-hardness of computing the objective function under Independent Cascade (IC) model, we devise an algorithm called Complementary Maximum Coverage (CMC) based on analyzing the influence of the nodes over the groups, ensuring the task of maximizing the number of activated groups. In addition, we also propose an algorithm called Improved Reverse Influence Sampling (IRIS) via adjusting the famous Reverse Influence Sampling (RIS) algorithm for GIM. Lastly, experiments are carried out to demonstrate that our CMC and IRIS both outperform the known baselines including Maximum Coverage and Maximum Out-degree algorithms in the average number of activated groups under IC model.