融合表情符号动态特征的舆情分析研究
2021-08-06李楠张羽卉
李楠 张羽卉
摘 要:[目的/意義]表情符号在网络舆情分析中的作用和价值逐渐显现,然而复杂网络环境下表情符号呈现出语境依赖、情感极性多元化等动态化特征,对于表情符号动态变化规律的研究不仅有利于丰富对表情符号使用规律的认识,更有助于提升舆情分析的准确度和有效性。[方法/过程]本文通过对微博评论中表情符号使用分布、情感异化等的多角度分析,挖掘表情符号的动态变化规律并在此基础上建立融合表情符号动态特征的网络舆情倾向分析模型。[结果/结论]实验证明:表情符号分布特征和情感变化规律具有显著的主题相关性、极性相关性及情感异化等动态特征,将其引入网络舆情分析能有效提升情感分析的识别精度。
关键词:表情符号;网络舆情;情感分析;动态特征;微博评论;舆情倾向分析
DOI:10.3969/j.issn.1008-0821.2021.08.010
〔中图分类号〕G206;TP391 〔文献标识码〕A 〔文章编号〕1008-0821(2021)08-0098-11
An Analysis of Web Opinion Combining the Dynamic Characteristics of Emoji
Li Nan Zhang Yuhui
(Institute of information science and technology information,East China University of Science and
Technology,Shanghai 200237,China)
Abstract:[Purpose/Significance]The value of emoji in the analysis of network public opinion gradually appear,but emoji in complex network environment presents dynamic characteristics such as context dependence and emotional polarity diversification.The study of the dynamic change law of emoji is not only helpful to enrich the understanding of the use of emoji,but also to improve the accuracy and effectiveness of public opinion analysis.[Method/Process]Based on the multi-angle analysis of emoji usage distribution and emotion change in Weibo comments,this paper excavates the dynamic change law of emoji and establishes the web public opinion tendency analysis model which integrates the dynamic characteristics of emoji.[Result/Conclusion]The comparative experiment of Weibo public opinion tendency analysis based on emoji shows that the regular change of emoji expression emotion has a positive effect on the analysis of public opinion tendency,and the introduction of emoji dynamic characteristics can effectively improve the recognition accuracy of emotion analysis.
Key words:emoji;web opinion;emotion analysis;dynamic characteristics;Weibo comments;analysis of tendency of public opinion
相比于文本语言,表情符号能够承载更多的情绪特征,传达丰富的语义信息,其符号化的沟通方式不仅丰富了网络语言,改变了人们的表达习惯和交流方式,也逐渐成为研究者们分析社交媒体观点、情感的重要特征之一。然而在不同类型语境中表情符号表达的情感并不是一成不变的,例如[微笑]表情,可以表示欣慰、认同等积极含义(例如:欢迎回家,你们辛苦了),也会被引申为失望、愤怒等消极负面含义(例如:“看了只想生气”),这种情感的动态变化给网络舆情的倾向分析带来了不确定性。……
