An Individual-differences-based Method for Discovering Viewpoints on Interactive Behavior

In this paper, we proposed an interactive behavior model for opinion tolerance and relationship difference. Furthermore, we used the tolerance measurement method based on trust mechanism to calculate the tolerance difference of individual opinion interaction, and constructed an individual opinion interaction behavior model based on attribute difference. This model is extended to social networks with relationship differences, and the occurrence probability of opinion interaction behavior is obtained by normalization method, so as to fit the change of real individual opinion and the polarization process of group opinion. We explored sociologically the corresponding relationship between parameter changes in the model and social phenomena. According to the specific needs, the optimization objective of the model is determined, which is solved by non-convex optimization method. Experiments show that the model can accurately describe the change of individual opinion and the process of polarization of group opinion, and can be applied to the non-polarization intervention of Internet opinion.