基于网络社群的网络舆情演化分析及突发事件预警机制研究

The administration of network public opinion is one of the vital components associated with cultural construction and innovative social management in the new era. Therefore, it is an urgent project to get comprehensive understanding of public opinion evolvement on the internet and construct an early-warning mechanism, which is crucial to the regional lasting stability and durable peace. While network events and Netizen's opinion develop in the network community, there is a strong probability of emergent collective behavior. This project is aimed to build the system of Internet public opinion evolvement and to enhance the efficiency of opinion analysis and early-warning on the basis of social network and event network. The research work will first deal with the hotspot discovery in network community and then analyze law of public opinion spreading within and between network communities by link prediction. Furthermore, in order to examine and forecast the evolving trend of network events, the dynamic development of network communities will be revealed, and by which the dynamic evolving models of public opinion will be established. Meanwhile, the project will combine the network community and cloud computing, build public opinion mining interface by the means of cloud statistics and thus enhance the efficiency of opinion analysis and intermediate computations storage. Finally, early-warning mechanism of public opinion based on human-computer collaboration will be constructed to put a timely halt to the transformation of opinion tendency into collective behavior, by which the anticipation efficiency of emergencies are to be greatly improved and powerful decision support can be provided to the government.

网络舆情管理是新时期文化建设和创新社会管理的重要内容之一，把握网络舆情发展态势和建立突发事件预警机制是地区长治久安亟待解决的课题。网络事件和网民观点在网络社群上不断演化，容易诱发群体行为突发事件。本课题在社会用户网络和舆情事件网络的基础上，构建网络舆情演化体系，提高舆情分析和和预警的效率。主要探讨网络社群内的热点主题发现，利用网络链接预测方法分析舆情信息在社群内和社群间的传播规律。揭示网络社群的动态演化规律，建立基于网络社群的网络舆情动态演化模型，分析与预测网络事件的演化趋势。将网络社群与云计算结合起来，建立基于云统计的舆情挖掘平台，提高舆情分析的时间效率和中间计算结果的存储效率。建立人机协同的网络舆情预警机制，及时遏止舆情事态向群体行为转化，增强突发事件预判效能，为政府相关部门提供决策支持。

网络舆情管理是创新社会管理的重要内容之一，本课题在社会用户网络和舆情事件网络的基础上，构建网络舆情分析体系;利用社会网络方法分析揭示网络舆情信息的传播规律，探讨网络热点主题下的网民行为分析；揭示了基于网络社群的舆情动态演化规律，分析了网民在线兴趣行为的演化趋势，对微博网民的网络影响力进行统计挖掘。将网络社群统计聚类与云计算结合起来，建立舆情挖掘统计分析平台，提高云统计的计算效率。利用共现矩阵对网络舆情态势进行分析，对网络舆情情感倾向性进行研究，初步建立人机协同的网络舆情预警机制，对新疆网络舆情预警提出相关建议。项目为新疆网络信息管理提供舆情技术服务，为新疆社会管理和地区长治久安提供决策知识支持。