Dynamic Social Influence Modeling from Perspective of Gray-scale Mixing Process

Social factors are useful in information and communication research. Researchers have recently been trying to utilize people's social factors on many topics, such as those regarding recommendation systems, decision making, and behavior predictions. However, they have mainly focused on estimating final results of people's decisions or actions, and few of them have ever considered median processes that explain how people's attitudes would change. Furthermore, some realistic factors and questions, such as interactions between people and people and unequal relationships in social ties, that widely exist in our common lives and have significant impacts on attitudes and that influence processes have rarely been well considered. In this paper, we propose a novel way of modeling dynamic attitudes changing on the basis of people's social structures. We defined and used different parameters to test and then validate our ideas. We also compared the results from a method of machine learning and our proposed model. In conclusion, we described why our proposed model had high levels of scalability to suit different and complex social influence cases.