Dynamic public opinion evolvement modeling and supervision in social networks

Abstract

With the booming of social networks, a large proportion of public opinion is expressed and transferred through social networks. With complex structure and varied evolving patterns, monitoring public opinion in social networks is not well solved for a long period. Motivated by the purpose of public opinion and social network evolvement rules, we developed a public opinion dynamic evolvement model and supervision mechanism in social networks. We assume our research target is a topic-based and opinion-driven social network that is the most popular one in studying public opinion. The background network of our model is a temporary social connection that we name as tornado-type social network (TTSN). In a TTSN, public opinion evolvement is decided by two basic factors: sentiment activity (SA) and opinion consistency (OC). Based on the observation of SA and OC, we have designed a model to supervise and optimize the public opinion express in social networks. Under the model, the public opinion supervision is regressed to an optimization problem. By solving the problem, both our deduction and simulation results show that public opinion in a social network tends to evolve from chaos to consistency, and SA follows approximately ideal normal distributions before a time limit T.