Maximizing positive influence in competitive social networks: a trust-based solution (Q2055555)

This paper studies positive influence maximization in competitive social networks. A trust-based solution is proposed, which stresses on the trust relationship in online social networks. The model leverages on the competitive influence diffusion which models the dynamics of the influence spreads. Positive influence is modeled by trust and negative influence is modeled by distrust. To compute the influence probabilities in the trust-based competitive influence diffusion, the paper uses the generalized network flows methods. The paper also proposes a trust-based competitive influence maximization algorithm to derive the positive marginal gain for users who are not seeds. The results are tested on some realistic datasets and the experiments show good efficiency of the proposed method.