基于半监督学习的微博情感倾向性分析

被引：5
作者：
朱玺
董喜双
关毅
刘志广
机构：
[1] 哈尔滨工业大学计算机科学与技术学院
来源：
山东大学学报(理学版) | 2014年 / 49卷 / 11期
关键词：
情感分析; reserved self-training; 训练度阈值;
D O I：
暂无
中图分类号：
TP393.092 [];
学科分类号：
080402 ;
摘要：
微博情感倾向性分析通常指对中文微博中每个句子褒义、贬义或者中性的情感进行自动分类。针对微博碎片化和情感类别失衡的特点,在半监督学习reserved self-training方法的框架基础上提取了适用于微博情感分类的文本特征,并提出了针对情感倾向性分析通过训练度阈值设定的方法来优化reserved self-training迭代终止的条件,在保留reserved self-training能有效处理微博语料中语料情感不平衡问题的优点基础上,防止了训练过度情况的发生。COAE 2014微博情感倾向性评测结果证明了该方法的有效性。
引用
收藏
页码：37 / 42
页数：6
相关论文
共 11 条
[1]  
Modeling public mood and emo-tion:Twitter sentiment and socio-economic phenomena. Johan B,Alberto P,Huina M. IC-WSM11 . 2011
[2]  
Occam’s Razor. Anselm Blumer,Andrzej Ehrenfeucht,David Haussler,Manfred K. Warmuth. Information Processing Letters . 1987
[3]  
Semi-supervised learning for imbalanced sentiment classification. LI Shoushan,WANG Zhongqing,ZHOU Guodong,et al. Proceedings of International Joint Conference on Artificial Intelligence (IJCAI) . 2011
[4]  
Robust sentiment detection on Twitter from biased and noisy data. Barbosa L,Feng J. Proceedings of the 23rd International Conference on Computational Linguistics . 2010
[5]  
Sentence level sentiment analysis in the presence of conjuncts using linguistic analysis. MEENA A,PRABHAKAR T. . 2007
[6]  
Sentence-level subjectivity detection using neuro-fuzzy models. RUSTAMOY S,CLEMENTS MA. Proceedings of the4th Workshop on Computational Approaches to Subjectivity,Sentiment and Social Media Analysis . 2013
[7]  
Sentiment analysis of political tweets:towards an accurate classifier. BAKLIWAL A,FOSTER J,VAN DER PUIL J,et al. Proceedings of NAACL Workshop on Language Analysis in Social Media . 2013
[8]  
Reserved self-training:a semi-supervised sentiment classification method for Chinese Micro-blogs. LIU Z,DONG X,GUAN Y,et al. Proceedings of IJCNLP . 2013
[9]  
Sentiment analysis on Chinese words and sentences based on maximum entropy model. DONG X,GUAN Y,LI B,et al. Proceedings of COAE . 2009
[10]  
网络评论如何影响网络购物意愿?[J]. 王远怀,于洪彦,李响.  中大管理研究. 2013(02)
12