Maximizing the Spread of Positive Influence Under LT-MLA Model

Since the steady development of online social networks, it has resulted in wide-spread research works in the social network analytical areas, especially in the area of influence maximization. Previous works study the influence propagation process based on the social influence model. But traditional influence models ignore some important aspect of influence propagation. The drawbacks of the models are the simplistic influence diffusion process, and the models lack attitude states and capability to capture the interaction between users. To address these problems, we modify Linear Threshold model based on multi-level attitude and users’ interaction, which is proposed modeling the positive and negative attitude towards an entity in the signed social network and the effect of interaction relationship between users. Then we propose the LT-MLA greedy algorithm to solve the positive influence maximization problem. Finally, we conducted experiments on three real-world data sets to select initial k seed with the positive attitude. The results show that the proposed solution in this paper performs better than other heuristic algorithms.