基于Agent和K核分解的群体事件微博传播模型

Agent and K-Shell based Diffusion Model of Mass Incidents in Weibo System

【目的/意义】有别于众多现有的针对社会群体事件的定性理论研究,文章采用了定性分析和定量研究相结合的方式探讨群体事件在微博网络中的传播特点,并给出了社会群体事件在微博中传播的传播模型。【方法/过程】以定性分析和定量讨论相结合的方式为指导思想,以复杂系统的理论为依据,提出对群体事件的微博传播进行基于Agent的建模分析;以复杂网络的相关知识为支撑,提出依据K-Shell index方式来划分Agent群类并设计了相应的交互规则。【结果/结论】实证分析了以K-shell index为依据构建Agent的优点;基于Repast平台对所设计的模型进行了仿真分析、模拟出了由个体用户的意见从无序交互到集中爆发的涌现过程,并以真实群体事件的数据为基础论证了该模型的科学性和实用性。这些工作能帮助人们更科学地理解群体事件微博传播中的演变规律,从而为群体事件的及时预测和有效应对提供理论指导。

[Purpose/siguificance] Most of available papers focusing on mass incidents in China are just pure ＂theoretical research＂. In order to have one more step to deeply understand the mass incident diffusion, beyond theoretical discussion, we choose to do analysis based on real Weibo data from previous incidents, and then we come up with the diffusion model. [Method/process] This paper follows the guiding ideology with the combination of qualitative analysis and quantitative discussion. Based on the theory of Complex System, we explain that Multi-agent simulation is a better method to investigate such kind of diffusion. Based on the theory of Complex Network, we point out that K-Shell index method is an effective way to identify different agent and it can help people on better simulation. [Result/conclusion] We come up with the diffusion model of mass incidents in Weibo system and explain why K-shell decomposition is helpful to this agent model. Based on this model, we successfully simulate the diffusion of mass incident in Weibo. Furthermore, according to the real data from famous mass incident, we show the effectiveness and efficiency of this agent model. We do believe this method can make more sense for people to understand the diffusion mechanism of mass incidents and come up with better strategy to manage them.