摘 要: 微博作为一种用户发表看法和观点的载体已成为互联网上一个重要的情感交流平台,博文搜索为这种交流 提供了方便快捷的途径。基于HowNet等中文情感词典的微博情感词的抽取和分类,计算词语语义相似度和倾向性。对文 本情感倾向的加权值、表情、和情感词增强因素等进行综合考虑。实验结果表明表情情感倾向对微博情感倾向起着重要作 用;在表情和文本情感倾向比值固定的情况下,调整因素和中性区间的选择会对情感倾向判断准确率产生影响;通过与基 于HowNet语义相似度的计算模型比较,该文方法使得情感倾向判断准确率有所提高。 |
关键词: 情感提取;情感分析;微博文本 |
中图分类号: TP399
文献标识码: A
|
|
An Analysis of the Emotional Tendency Based on Micro-Blog Text |
SONG Jihong,GE Daming
|
( School of Information Science and Engineering, Shenyang University of Technology, Shenyang 110870, China)
|
Abstract: As a carrier for users to express opinions and viewpoints,micro-blog has become an important platform for emotional communication.The function of blog search brings more convenience and efficiency to the communication. The paper proposes an algorithm to extract and categorize the emotional words in micro-blog based on HowNet and other Chinese dictionaries of emotional words.Comprehensively considering weight values of text emotional tendency,emoticons and enhancement factors of emotional words,the algorithm computes semantic similarity and tendency among words and expressions.The experiment result shows that:the emotional tendency of emoticons plays an important role in the emotional tendency in micro-blog,the selection of adjustment factors and neutral ranges affects the judgment accuracy of emotional tendency under the circumstance that the emotional tendency ratio between the emoticon and the text is fixed,compared with the algorithm model based on HowNet semantic similarity,the method proposed in this paper effectively improves the judgment accuracy of emotional tendency |
Keywords: sentiment extraction;emotion analysis;micro-blog text |