Analysis of Cyber Interactive Behaviors Using Artificial Community and Computational Experiments

An artificial community is constructed for studying cyber interactive behavior of message publishing by considering impacts from individual's psychological state and social environment. The bottom-up multiagent approach is applied to model the interactions between individuals, and the evolution and organization of collective cyber behaviors. The cyber message publishing behaviors in response to online topics is demonstrated in the artificial community, where each agent perceives the state of the topics and make decisions to operate on specific themes with particular message publishing mechanisms. The mechanisms are designed in line with the analysis of statistical results of 10 years data from Tianya.cn, which are further validated through the artificial community using computational experimental leverages. Experimental results illuminate that the individual's post publishing behavior is dominated by agent's psychological states, while the comment publishing behavior is more susceptible to the agent's social environment expressing the collective efforts of topics and other agents in the community. The artificial community has been used to test the impact of several factors (e.g., individual's psychological state, number of initial agents, and number of new added agents) to individual's message publishing mechanism and to cyber collective behavior patterns. It provides a software-defined social laboratory for social and cyber behavior studies with computational experimental measures.