网络舆情态势分析研究

Trend analysis of network public opinions

对网络舆情的分析已引起人们的高度重视,因为网络舆情热点具有导向性作用,如"某某门"事件,在短短60小时过后就引起监察部门的注意,这只有在日益流行的网络应用的今天才可能做到.分析网络舆情热点的手段多种多样,角度不同,如:可利用文本聚类的方法、动态文摘生成的方法以及意见领袖特征等方法来发现网络舆情中的热点和焦点信息.截止目前,分析网络舆情热点的方法没有固定的模式,值得深入探讨和研究.针对目前网络舆情研究的典型方法及其思想,进行概述性分析,给出网络舆情分析中词频、权重特征词、意见领袖属性矩阵等概念的形式化定义.本文作者通过对比分析,比较研究了网络舆情方法的优缺点;总结了国际上在这一领域的最新研究进展和研究方法,给出下一步的研究工作重点.

Due to the guiding role of the hot Internet public opinions, it attaches great importance to ana-lyze the Internet public opinions. Taking the Pornographic event as a typical event, it will not give rise to the attention of the inspection and supervision departments in only over 60 hours without the Web. There are several approaches to find the focus information of Internet public opinions, for example, text clustering, dynamic abstracts generating and opinion leader. Currently, the methods of analyzing the hotspot of Internet public opinion have no fixed models. It is still worth exploring and doing research on this popular research area. This study provides overview analysis of several research approaches and ide- as on Internet public opinions. It gives a formal definition of the concept of word frequency, weighting feature words, attribute matrix of opinion leader. It compares the typical research methods on Internetpublic opinions, addresses the limitation of these works, and shows the overview of research methods and progress in this field all around the world. Finally, it addresses the focus of future research.