摘要
情感计算是对文本进行情感分析与挖掘,得出文本评价调性的过程。基于知网的情感计算方法首先依据知网"情感分析用词语集"构建情感词典,然后计算文本中特征词汇的情感倾向,最后通过累加得出文本的情感倾向。实验结果表明,在设定一个合理中性阀值后,能较好地计算出特征词汇和篇章的情感倾向,有较高的查全率和查准率。
Affective computing is an evaluation process, which analyzes the sentiment of the text. The paper discussed the affective lexicon based on Hownet. And the lexicon is used to measure the feature words' se- mantic orientation. Text' s semantic orientation could be obtained by accumulating. Experimental results show that the method has a higher recall and precision rate
出处
《河南城建学院学报》
CAS
2012年第3期30-33,共4页
Journal of Henan University of Urban Construction
基金
安徽省教育厅重点研究项目(20101689)
关键词
情感计算
知网
情感词典
语义相似度
affective computing
Hownet
affective lexicon
semantic similarity