文本分类方法在网络舆情分析系统中的应用研究  

Applications of Text Classification in Network Public Opinion System

作　　者：马海兵[1] 毕久阳 郭新顺[2] 

机构地区：[1]武警政治学院基层政治工作系,上海200435 [2]上海对外经贸大学商务信息学院,上海201620

出　　处：《情报科学》2015年第5期97-101,共5页Information Science

基　　金：2010年度国家社会科学基金项目(10CTQ024);上海市高校085基金项目(Z08511030)

摘　　要：针对文本分类技术在网络舆情分析系统中的应用展开研究。介绍了文本分类的基本原理;实现了KNN和SVM两种分类算法;结合实际项目,展示了文本分类方法在网络舆情主题分类、主题检测与跟踪及舆情观点倾向性分析中需要解决的具体问题,给出应用实例。实验结果给出了KNN和SVM分类方法的微平均查准率,表明了这些方法的有效性。Applications of text classification in network public opinion system are addressed in this paper. Firstly, the basic principle of text classification is introduced. Secondly, two classification methods, KNN and SVM are established. Thirdly, combined with the actual project, these methods are used in net- work public opinion system, demonstrating the need to address the specific problems, such as implement- ing subject classification, topic detection and tracking, and orientation analysis of public opinion. Eexper- imental results are given about micro-average precision of KNN and SVM, indicating the effectiveness of these methods.

关 键 词：舆情分析 文本分类 主题检测与跟踪 倾向性分析 

分 类 号：G250.2[文化科学—图书馆学] G254