摘要
针对目前关于突发事件的网络舆情监测研究中缺乏连续动态监测的方法的问题，提出了基于扫描统计量的微博中突发事件舆情监测方法。本文首先对微博中的网络舆情要素进行分析，提出了基于用户信息量和微博项信息量的计量方法。进而，提出了基于时间扫描统计量的微博中的网络舆情监测方法，具体包括时间扫描变量及概率分布模型确定、微博数据时间扫描窗口定义、时间扫描窗口对数似然比计算、对聚集窗口的非随机性进行显著性分析。最后，通过实验对比验证了该方法的有效性。解决了目前在突发事件相关的网络舆情监测中舆情爆发时间点无法进行有效判断的问题，同时为网络舆情的监测研究提供了一个连续动态监测的方法。
For the lack of continuously dynamic monitoring method in Emergency related public opinion monitoring researches, we propose a Scan Statistic-based Emergency related public opinion monitoring method in Micro-blogs. Through analyzing the elements of network public opinion in microblog, a quantitative method is proposed based on the information volume of users and Micro blog entries at first. And then, the time scan statistics-based emergency network public opinion monitoring method is put forward, which consists of the determination of time Scan statistics and the probability distribution model, the definition of Micro-blogs data time scanning window, the calculation of logarithmic likelihood ratio of time scanning window and the non-randomized significance analysis of cluster window. In the end, we verified the effectiveness of the proposed method by related experiment. This research solved the existing problem of effectively determining the public opinion outbreak time in network public opinion emergency monitoring, ana provides a method of continuously dynamic monitoring for Network public opinion monitoring research.
出处
《情报学报》
CSSCI
北大核心
2015年第4期414-423,共10页
Journal of the China Society for Scientific and Technical Information
基金
中央高校基本科研业务费专项资助项目(DUT14RW209)
国家自然科学基金(71073017)研究成果之一
关键词
网络舆情
舆情监测
突发事件
微博扫描
统计量
network public opinion, public opinion monitoring, emergency event, micro-blog, scan statistic
