Abstract:

To help people have a better understanding of the mechanism of message dissemination in micro-blog when mass incident happens and have a better idea of how to identify the key users in the diffusion network, the micro blog of social group events is analyzed based on the real case communication data by combining qualitative analysis and quantitative research. Besides, a micro blog communication system of group events is constructed through the information forwarding relationship of events. Moreover, a comprehensive evaluation model of the importance of group event user communication is constructed through the principal component analysis of some basic communication indicators. Furthermore, a correlation analysis of the evaluation value of the model and the central indicators of complex network is carried out in combination with the related theory of complex network. The analysis and research indicate that the network of group events is a complex network with a scale-free nature, small world characteristics, and a high clustering coefficient, in which, the evaluation value of user communication importance is highly correlated with the K-core decomposition index of the user complex network. Therefore, people can rely on the K-core decomposition analysis method to find more accurate and effective important users who affect group event micro blog communication.

Key words: mass incidents, complex network, diffusion in micro-blog, principal component analysis, centrality

摘要：

为了达到更进一步理解群体事件在微博中的传播机理并得到能够综合评定事件参与用户在传播中重要程度的方法，采用定性分析和定量研究相结合的方式，基于真实案例的传播数据对社会群体事件的微博进行了分析。通过事件的消息转发关系构建了群体事件的微博传播系统；通过对若干传播基础指标实施主成分分析，构建了群体事件用户传播重要性的综合评价模型；结合复杂网络的相关理论，开展了该模型的评价值和复杂网络中心性指标的相关性分析。通过分析和研究，发现了群体事件的网络是具有无标度性质、小世界特征和较高集聚系数的复杂网络，在该网络中用户传播重要性的评价值实际与用户的复杂网络 核分解指标值存在极高的相关性。因此，人们可以依靠 核分解为核心的分析方法更精确、更有效地找到影响群体事件微博传播的重要用户。

关键词: 社会群体事件, 复杂网络, 微博传播, 主成分分析, 中心性指标