With the worldwide proliferation of social networks, public opinion analysis of data generated by social networks has become an important field of research. Social networks have become a major platform for public opinion formation and diffusion, and analyzing public opinion through social network data plays an important role across numerous fields, including political science, economics, commerce, finance, international trade, public policy implementation and so on. Nevertheless, the corresponding quantitative indexes of public opinion analysis have not yet been developed, and the theoretical foundation underpinning such indexes has yet to be established. How to measure public opinion through social network data is a significant problem in need of the development of a series of quantitative assessment indices and social computing methods that can be used to solve this problem. This paper proposes both the concept of a public psychological pressure index and its calculation method, making it a fundamental work in the field of public opinion analysis. The maximum entropy principle is introduced to the social computing domain in this paper and positions it as the theoretical foundation underpinning such indexes.INDEX TERMS Information entropy, intelligence and security informatics, public information security, public opinion analysis, public psychological pressure index, social computing.

Microblog is a micromessage communication network in which users are the nodes and the followship between users are the edges. Sina Weibo is a typical case of these microblog service websites. As the enormous scale of nodes and complex links in the network, we choose a sample network crawled in Sina Weibo as the base of empirical analysis. The study starts with the analysis of its topological features, and brings in epidemiological SEIR model to explore the mode of message spreading throughout the microblog network. It is found that the network is obvious small-world and scale-free, which made it succeed in transferring messages and failed in resisting negative influence. In addition, the paper focuses on the rich nodes as they constitute a typical feature of Sina Weibo. It is also found that whether the message starts with a rich node will not account for its final coverage. Actually, the rich nodes always play the role of pivotal intermediaries who speed up the spreading and make the message known by much more people.

No abstract

No abstract

Based on the modification of the convergence parameter μ in the Weisbuch-Deffuant (WD) model, we investigated the influence of the network structure on opinion dynamics by comparing the processes of opinion dynamics in the Watts-Strogatz (WS) small-world network and Barabási-Albert (BA) scale-free network. The simulation results present that the time evolution of opinions does not always end up with a consensus; the final number of opinion clusters depends on the value of the bounded confidence but compared with the situation in the original WD model the effect of the bounded confidence is different when dynamics happens in the small-world network and BA network. Furthermore, the structural cohesion of the network is strengthened by the rich nodes in the BA network, which make the opinions evolve at a much faster rate than those in the small-world network.