摘要
不同语境的词语其倾向性有差异。针对此问题,研究了二义词倾向性识别,词语倾向性识别及评论倾向性分析。建立了结构模板集和核心词模板集,用模板匹配方法判断词语在特定句子中的倾向性。根据词语的倾向性并考虑副词和反问句对倾向性的影响,计算评论的倾向性。实验结果表明,可提高网络评论倾向性分析的准确率。
Words have different orientations in different contexts. In order to solve the problem, ambiguous word orientation identification, and a series of work such as word orientation identification and Web review orientation analysis are studied. The sets of structure template and keyword template are built, and word orientation is analyzed in the specified sentence through template matching. At last, the orientation of Web review is analyzed according to the orientation of the words and the impact of adverbs and rhetorical questions. Experimental results show that the accuracy of Web review orientation analysis is improved.
出处
《计算机工程与应用》
CSCD
2012年第32期136-139,199,共5页
Computer Engineering and Applications
基金
河南省科技攻关计划项目(No.102300410198)
关键词
词语倾向性
结构模板
核心词模板
评论倾向性
word orientation
structure template
key word template
web review orientation