语义规则在微博热点话题情感分析中的应用  

Application of semantic rules to sentiment analysis of microblog hot topics

作　　者：赵文清[1] 侯小可[1] 沙海虹 

机构地区：[1]华北电力大学(保定)控制与计算机工程学院,河北保定071003 [2]英业达集团(北京)电子技术有限公司开发部,北京100086

出　　处：《智能系统学报》2014年第1期121-125,共5页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金资助项目(70671039);中央高校基本科研业务费专项资金资助项目(12MS121)

摘　　要：近来,针对微博热点话题的情感分析研究得到了广泛关注,而基于监督的学习方法在分析文本时会忽视词语的上下文联系。根据中文微博的特点,提出了一种基于语义规则的方法对微博热点话题进行情感分析。该方法首先需要人工整理出程度副词表、否定词表和微博中默认表情符号的褒贬分类。然后在情感词语计算的基础上,考虑上下文中否定词和程度词对修饰情感词语的情感倾向和情感强度的影响,同时也设定规则计算表情符号对一条微博的情感倾向判断的作用。最后与基于情感词典的方法做实验对比,实验结果表明该方法在文本情感倾向性识别的准确率上有了一定提高。The research on the sentiment analysis for microblog hot topics has attracted much attention recently , while the studying method on the basis of supervision neglects the context of a word in the analysis of text .Accord-ing to the characteristics of Chinese microblogs , a method based on semantic rules is proposed for sentiment analy-sis of microblog hot topics .As for the method , firstly, we need to manually sort out a degree adverb list , a negative word list and the appraisal category of the expression symbols defaulted in a microblog .Secondly , on the basis of the calculation of sentiment words , we consider the impact of negative words and degree words in the context of the emotional tendency and strength decorating sentiment words;in addition, we also set rules for calculating the influ-ence of the expression symbol on the sentiment tendency judgment of a piece of microblog .Finally, our proposed method is compared with the method based on the emotional dictionary .The experimental results show that the pro-posed method improves the identification accuracy of the text sentiment tendency .

关 键 词：微博 热点话题 情感分析 语义规则 情感词典 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]