Maximizing influence in social networks by distinguishing the roles of seeds

The problem of influence maximization in social networks has attracted much attention, and some algorithms have been proposed. However, the existing methods may be not suitable for large-scale networks because of high time complexity or failing to achieve the great performance. Given that the impacts of seeds in the loose neighbors (i.e., only including one-hop path with the candidate node) and the close neighbors (i.e., including one-hop and two-hop paths with the candidate node, and they form a closed triad) on the degree discount of candidate node are different. Moreover, when selecting multiple nodes as the seeds, we should not only consider the importance of the seeds themselves, but also ensure that the seeds are sufficiently dispersed to avoid the redundancy of propagation. To the end, we propose an efficient heuristic algorithm for influence maximization in social networks by considering redundancy weakening and two types of seeds into degree discount (named RWTDD). Based on the independent cascade model, the proposed RWTDD is compared with some well-known heuristic algorithms and greedy algorithms in six real social networks, experimental results indicate that the proposed RWTDD has a better influence coverage and low time complexity.