摘要
目前对微博评论的研究主要聚焦在影视、购物等非电力领域,而对电力领域的研究相对较少.因此在影视等领域的研究基础上,根据电力行业的特性,将评论进行预处理后,建立评论关系树,使用动态扩展情感词典和基于支持向量机的方法,建立情感极性判别规则,进行情感极性分析.经实验验证,生成评论关系树后,扩展情感词典和支持向量机两种方法在电力领域的正确率均得到了明显的提升.
Presently,the research on microblogging is mainly focused on non-power fields such as film and television,and research on power industry is relatively limited. Therefore,based on the research of the field of film and television,according to the characteristics of the power industry,after the pretreatment of the comment,the relationship tree is set up,and by using the dynamically extended emotional dictionary and based on support vector machine method,the emotional polarity discrimination rules are established to carry out emotional polarity analysis. After experimentally verifying and generating the relationship between the tree,the extended emotional dictionary and the support vector machine improve the correctness rate of the two methods in the power field.
出处
《上海电力学院学报》
CAS
2017年第6期601-606,612,共7页
Journal of Shanghai University of Electric Power
关键词
情感分析
微博评论
评论关系
扩展情感词典
支持向量机
emotional analysis
microblog comment
comment structure
extended emotional dictionary
support vector machine