基于聚类分析的网络舆情倾向性分析研究

Analysis of Tendentiousness of Internet Public Opinion Based on Cluster Analysis

介绍了国内外网络舆情分析的理论研究、系统构建和关键技术研究现状,给出了基于聚类分析的网络舆情倾向性分析的距离模型和相关系数模型,提出了基于时间片的k中心点法聚类分析算法,给出了该算法实现网络舆情倾向性分析的流程;以论坛和微博作为实验数据抽取平台,按照时间片进行信息的随机抽取,验证了选择5个连续的时间片且针对不同的样本数的聚类分析结果,能有效反应出网络舆情演化的倾向性,降低了聚类的维数,增加了聚类的可靠性;验证了本文提出的模型能有效地提高舆情演化主题提取的查全率,较大幅度提高了时间效率,为网络舆情信息倾向性分析提供了有益的解决方案。

This paper introduced the theory of Internet public opinion analysis,the construction of department and current situation of key technologyresearchat home and abroad.The distance model and correlation coefficient model of Internet public opinion analysis based on cluster analysis were given.K-mediods method clustering algorithm based on time slice was proposed,and the process was given.Using forum and micro-blog as experimental data extraction platform,we verified the clustering results of 5 continuous time slices for different samples.K-mediods method clustering analysis can effectively reflect the tendency of Internet public opinion evolution,reduce the dimension of clustering,and increase the reliability of clustering.It is verified that the method can improve the recall rate and greatly improve the time efficiency.It provides a solution for the tendentiousness analysis of Internet public opinion.