Opinion influence maximization problem in online social networks based on group polarization effect

Group polarization effect reveals a phenomenon that discussions could lead to an increase in the extremism among members in a group. For the first time, the information propagation mechanism of group polarization effect is detected and modeled by real-world data in this paper. In addition, the adverse impact of opposing opinions is considered, which well portrays the reality that negative reviews in viral marketing could damage reputations and long-term profits. The realistic dynamic enhanced independent cascade (EIC) model is designed based on group polarization effect and individual preferences. The problem is called the effective influence maximization with group polarization effect (EIMGP) problem, where individuals are selected to post information while maximizing the positive impact of approval and minimizing the negative impact of disputes. Secondly, the objective function is neither submodular nor supermodular. Then, the group-based effective influence maximization (GEIM) algorithm is proposed to select seeds using network structures and model features. Finally, experimental results show that the GEIM algorithm could identify the group polarization effect, improve the quality of the seeds and reduce the running time.