An Influence Blocking Maximization Algorithm Based on Community Division in Social Networks

The spread of negative influences such as rumors and misinformation in Online Social Networks (OSNs) can threaten public safety. Therefore, the Influence Blocking Maximization (IBM) problem has received extensive attention in recent years. Although many researchers have investigated IBM problem, but there were some issues, including the imbalance between time consumption and performance in seed selection, as well as the influence overlap among selected seed nodes. In this paper, we present an IBM algorithm called IBM-CD based on community division to solve the IBM problem efficiently. This algorithm initially employs our proposed membership propagation approach to divide community centered around source nodes. Subsequently, the communities with low link strength will be merged. Additionally, we use the independent path to calculate the activation probability between node pairs, deriving the blocking effect on nodes based on this concept. Finally, we use the derived blocking effect as the metric to select positive seed nodes for influence blocking within each community. Through experimentation on real world datasets, we demonstrate that our proposed algorithm achieves faster and more effective suppression compared to existing methods.