摘要
[目的/意义]对国内文本情感分析的研究成果进行梳理与计量分析,有助于从不同角度了解主题研究状况、发文情况,对后续的研究具有一定的参考价值。[方法/过程]本文对发文期刊、作者团队进行统计分析,并利用共词分析法,研究关键词之间的联系,探讨近十年来在文本情感分析的研究热点及现状。[结果/结论]结果表明,我国近两年对文本情感分析的研究主要有基于情感词典的机器学习和神经网络的深度学习两种方法,文章最后指出了两种方法的研究现状及未来研究方向。
[Purpose/Importance]The research results of text sentient analysis were sorted out and quantitatively analyzed,which was helpful to understand the status of subject research and publishing,and it had certain reference value for subsequent research.[methods/Procedures]This paper made a statistical analysis of the published journals and author teams,and uses co-word method to study the relationship between keywords,and discussed the research hot spots and current situation of text sentient analysis in the past decade.[Results/Conclusion]The results showed that the research on text sentiment analysis in China in the past two years mainly included two methods:machine learning based on emotion dictionary and deep learning based on neural network.At last,the paper pointed out the future research direction of the two methods.
作者
陈红琳
魏瑞斌
张玮
张宇航
Chen Honglin;Wei Ruibin;Zhang Wei;Zhang Yuhang(School of Management Science and Engineering,Anhui University of Finance & Economics,Bengbu 233030,China)
出处
《现代情报》
CSSCI
2019年第6期91-101,共11页
Journal of Modern Information
基金
2015年安徽省高等教育振兴计划项目"信息管理与信息系统专业结构优化调整与专业改造"(项目编号:2015zytz021)
安徽财经大学科研项目"基于时间-关键词共词分析的大数据学科动态热点研究"(项目编号:ACKY1860)
安徽财经大学大学生创新创业训练计划项目"大数据推动社交网络个性化推荐的发展"(项目编号:201810378413)
关键词
文本情感分析
共词分析
情感词典
深度学习
神经网络
text sentiment analysis
co-word analysis
emotional dictionary
deep learning
neural network