网络舆情意见挖掘中极性词典构建和极性识别方法研究  

A Study on the Polar Word Identification Method and Polarity Dictionary Construction of Online Public Opinion Mining

作　　者：曹树金[1] 张学莲[2] 陈忆金[3] 

机构地区：[1]中山大学资讯管理学院,广州510006 [2]中国农业银行广州市分行东城支行,广州510500 [3]仲恺农业工程学院,广州510225

出　　处：《图书情报知识》2012年第1期60-65,共6页Documentation,Information & Knowledge

基　　金：广东省教育部产学研合作专项资金项目“网络舆情智能监测与分析系统”(2007A090302027)成果之一

摘　　要：本研究通过对网络新词的构造规律分析,对表达情感的网络新词进行识别算法的设计,提出基于HowNet和网络情感词的极性词典的人工构建方法,引进了同义词表减少手工构建的工作量,并抓住网络新词往往是现有情感词的其他形式这一特点,设计了针对表达情感的网络新词的自动识别方法,进一步扩展极性词的收录范围。Through the analysis of the construction law of new words from the internet, an algorithm was designed for identifying new words of network which express emotions. The artificial construction method of polarity dictionary on basis of the HowNet and the emotional vocabulary of the network was proposed, by introducing a synonym table to reduce the workload of manual building. Seiz- ing the characteristics that the new words of network are often the other forms of the existing emotional words, an automatic identifi- cation method of new words of network is designed, which further expands the scope of polar words.

关 键 词：舆情 意见挖掘 极性词 极性词典 极性识别 

分 类 号：G350[文化科学—情报学]