基于统计数据的微博表情符分析及其在情绪分析中的应用

发布时间：2018-06-28 16:45

  本文选题：表情符 + 情绪向量　； 参考：《计算机工程与科学》2016年03期

【摘要】：表情符作为一种新兴的网络语言,受到了越来越多的微博用户的青睐。微博中出现的表情符形象直观地表达了博主的情绪,对情绪分析起着至关重要的作用。首先对大量中文微博中表情符的使用特点、分布情况和情绪表达特点进行了统计分析。然后,人工选取具有代表性且情感倾向明确的表情符作为六类基本情绪的种子表情符。根据目标表情符和六类情绪的种子表情符在微博文本中的共现情况,为其建立六维情绪向量,并将其应用于微博情绪分析。在两个数据集上的实验结果表明,本文建立的表情符情绪向量有效地提高了微博情绪识别的精度。
[Abstract]:As a new network language, emoji is favored by more and more Weibo users. The appearance of emoji in Weibo intuitively expresses the blogger's emotion and plays an important role in emotional analysis. Firstly, the characteristics of emoji usage, distribution and emotional expression in Chinese Weibo are analyzed statistically. Then, the representative emoticons with clear affective tendency are artificially selected as seed emoticons of six basic emotions. According to the co-occurrence of the target emoji and the seed emoticons of six kinds of emotions in the Weibo text, the six-dimensional emotion vector is established and applied to the Weibo emotional analysis. The experimental results on two datasets show that the emoji emotion vector established in this paper can effectively improve the accuracy of Weibo emotion recognition.
【作者单位】： 南京航空航天大学计算机科学与技术学院;
【基金】：国家自然科学基金(61202132)
【分类号】：TP391.1;TP393.092
，

本文编号：2078674