摘要
随着互联网信息技术的发展,网络媒体已经成为民众表达情绪、意愿的主要场所,也是舆情信息传播的主要渠道.为了在海量的网络互动信息中能够准确迅速地对网民观点进行情感倾向性分析,本文提出了一种将句法分析和情感词典相结合的分析方法.首先,对目前通用的情感词典组织模型进行了改进,给出了具体的改进模型.然后,在用情感词典进行切词的基础上对句子进行语法分析,再利用情感词在句子中的成份、情感指数权重以及与其他情感词之间的组合共现关系计算出综合的情感指数.通过对论坛回复、新闻语评论内容的实际测试,验证了此方法的有效性.实验结果表明,此方法在一定程度上提高了情感分析的正确率,并具有一定的适应性和推广性.
With the Development of the Intemet and information technology, network media has become one of the main venues in public expression of emotions, and the main channel of public information dissemination. In the massive network of interactive infor- mation can be quickly and accurately to netizen opinions sentiment analysis, this paper presents a will of syntactic analysis and the method of combining sentiment dictionary. First of all, the general sentiment dictionary organization model is improved, given a spe- cific improvement model. Then, in the emotional dictionary segmentation based on grammatical analysis, and use of emotional words in the sentence composition, emotion index weights and other emotional word combinations between relation to calcu- late comprehensive emotion index. Through the forum replies, news-Comment on the content of the actual test, verify the validity of this method. The experimental results show that, this method can improve the correct rate of sentiment analysis, and has certain a- daptability and promotion.
出处
《小型微型计算机系统》
CSCD
北大核心
2014年第4期811-813,共3页
Journal of Chinese Computer Systems
基金
黑龙江省教育厅科学技术研究项目(12521055)资助
关键词
倾向性分析
情感分析
句法分析
情感词典
网络舆情
text orientation analysis
sentiment analysis
parsing
sentimental lexicon
network public opinion ]
作者简介
E-mail：13904861067@126．com肖红。女，1979年生，硕士，副教授，研究方向为数据挖掘与图像处理；
许少华。男，1962年生，博士，教授，研究方向为智能信息处理．
