群体性突发事件预警的超网络方法研究

项目名称： 群体性突发事件预警的超网络方法研究

项目编号： No.71473034

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 管理科学

项目作者： 王名扬

作者单位： 东北林业大学

项目金额： 52万元

中文摘要： 群体性突发事件的爆发严重威胁社会秩序和危害公共安全，认识事件的演化规律并实现及时准确的预警，对预防事件爆发具有重要意义。 群体性突发事件的舆情信息往往藏匿于海量的网络信息之中。作为一个复杂的信息传播系统，互联网包含着多种不同类型要素和要素之间错综复杂的交互关系。但是，现有研究多数局陷于单一要素模式网络中的信息传播研究。建立更为合理的、综合考虑互联网信息传播各要素的群体性突发事件信息传播模型，对实现更为可靠的群体性突发事件预警至关重要。 本项目在综合考虑群体性突发事件网络舆情各种维度信息的基础上，构建维度子网相互交织影响的超网络模型，作为描述群体性突发事件在线信息传播系统复杂构成和结构形态的基准模型；考虑到超网络模型的结构复杂性，发掘对事件发展态势起重要推动作用的核心拓扑结构；在此基础上，设计合适的核函数提取表征事件发展轨迹的网络演化模式特征；构建贝叶斯网络，实现对群体性突发事件的预警。

中文关键词： 群体性突发事件；预警；超网络；特征提取；贝叶斯网络

英文摘要： The burst of mass emergency has made great damages to social order and public security. In this case, figuring out the evolving laws of mass emergency and then making warning timely and accurately are important for predicting the burst of mass emergency. The public opinions of mass emergency are often hidden in the massive information of internet. As a complex information spreading system, the internet contains various kinds of elements and the complex interactive relationship among them. However, the current researches are mainly focusing on the information spreading based on one kind of element. A more reasonable information spreading model for mass emergency should be established which must consider the various kinds of elements in the information spreading process of internet. And this model is crucial for making reliable warning for mass emergency. By considering the various dimension information hidden in the network public opinion of mass emergency, this project construct a supernetwork which contains several subnets and the interactive relationship among them. This supernetwork would be used as the basic model for describing the complex formation and structure of the on-line information spreading system for mass emergency. Then the kernel topology structure of supernetwork is extracted out. This kernel topology structure plays vital roles on pushing the development of mass emergency. By designing feasible kernel functions, the network evolving characters could be obtained which map to the development tread of mass emergency. Finally, the Bayesian network is established by considering these characters, which could be used to make warn on mass emergency.

英文关键词： mass emergency;warning;supernetworks;feature extraction;Bayesian network