Improvement of Hegselmann-Krause Bounded Trust Model for Dynamic Evolution of Internet Public Opinion

Abstract

In the current network environment, the sources of information are gradually diversified and online public opinion events are often “enlarged” by the network, which will have an impact on network security, people's lives and social stability. So the prediction and monitoring of online public opinion has become an inevitable work. However, the majority of simulation studies on the evolution of public opinion, most scholars use social network analysis to study complex social networks. Such studies mainly consider discontinuous changes when selecting corresponding variables. However, the representation of variables in the evolution process is often overly idealistic, since the dynamic and continuous changes of the corresponding thresholds are not considered. The Hegselmann-Krause (H-K) bounded trust model offers a more comprehensive representation of the dynamic evolution of online public opinion. Still, the existing H-K bounded trust model has shortcomings in text content analysis. This paper presents the related research on the internet public opinion and its evolution, describes the H-K model and proposes an improved dynamic H-K model. Finally, a simulation experiment utilizing real-world data demonstrates the effectiveness of the improved dynamic model.