Adversarial influence maximization algorithm based on node local influence

In this paper, we propose a novel heuristic algorithm, the local influence algorithm (LI), to address the problem of maximizing adversarial influence in social networks based on the linear threshold model. First, we analyze the factors that diminish the influence of seed nodes and identify candidate nodes with specific topological features that can effectively block the propagation of seed nodes. Based on these features, we develop a new heuristic algorithm. Finally, we conduct experiments to perform a comparative analysis across different datasets and traditional methods. The experimental results demonstrate that the LI algorithm significantly outperforms baseline algorithms in both effectiveness and computational efficiency. Specifically, the blocking effect of the LI algorithm improves by an average of 88.64% compared to the best-performing baseline, Degree. Moreover, ablation experiments and tests on synthetic networks further validate the contribution of each topological feature in enhancing the blocking capacity of nodes within the LI algorithm.