一种基于情绪激励度的情绪词加权方法

J4 ›› 2016, Vol. 38 ›› Issue (02): 386-394.

一种基于情绪激励度的情绪词加权方法

王世泓,牛耘

（南京航空航天大学计算机科学与技术学院,江苏南京 210016）

收稿日期:2015-02-28 修回日期:2015-05-01 出版日期:2016-02-25 发布日期:2016-02-25
基金资助:
国家自然科学基金（61202132,61170043）；国家973计划（2013CD744904）

A weighting method of emotion words
based on the level of arousal

WANG Shihong,NIU Yun

（School of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China）

Received:2015-02-28 Revised:2015-05-01 Online:2016-02-25 Published:2016-02-25

摘要/Abstract

摘要：

在不同的上下文中，情绪词对情绪的激励程度会发生变化。现有情绪词典中大多数只标注了情绪词的情绪类别而未涉及情绪词的激励度。在极少数标注情绪强度的词典中,所标注的强度未考虑上下文的影响。提出一种根据上下文形成的情境评估情绪词对情绪的激励程度并据此对情绪词加权的方法。通过比较情绪词的共现模式与自身情绪类的分布模式计算情绪词的激励程度。然后根据激励程度计算情绪词的情绪权重并将其用于微博情绪识别。实验结果表明,与现有词典中的情绪强度相比,本文方法计算的情绪权重更准确地描述了情绪词在语料中表达的情绪,有效地提高了情绪分析的精度。并且本文方法还能够有效综合多个词典的优势,进一步提高微博情绪分析的准确率。

关键词: 情绪强度, 情绪词典, 语料上下文, 情绪激励度, 情绪权重

Abstract:

The level of arousal for an emotion word may vary in different contexts. However, this information is ignored in most emotion lexicons. We propose a weighting scheme of emotional words based on the level of arousal in various contexts. The level of arousal is evaluated by analyzing the cooccurrence patterns of emotion words as well as the distribution patterns of emotions. The weights of emotion words are calculated according to the level of arousal, and the results are used to weight emotion words in emotion analysis of microblogs. Experimental results on two data sets show that the proposed method outperforms current emotion lexicons. Moreover, the proposed approach can take advantages of multiple lexicons and further improve the accuracy of the system.

Key words: emotion intensity;emotion lexicon;corpus context;level of arousal;emotion weight

王世泓,牛耘. 一种基于情绪激励度的情绪词加权方法[J]. J4, 2016, 38(02): 386-394.

WANG Shihong,NIU Yun. A weighting method of emotion words
based on the level of arousal [J]. J4, 2016, 38(02): 386-394.

参考文献［15］

［1］	SemEval2007［EB/OL］.［20150220］.http://nlp.cs.swarthmore.edu/semeval/.
［2］	Ghazi D,Inkpen D,Szpakowicz S.Hierarchical versus flat classification of emotions in text［C］∥Proc of the NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text,2010:140146.
［3］	Bellegarda J R.Emotion analysis using latent affective folding and embedding［C］∥Proc of the NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text,2010:19.
［4］	Ekman.Facial expression and emotion ［J］.American Psychologist,1993,48(4):384392.
［5］	HowNet ［EB/OL］.［20150220］.http://www.keenage.com/.
［6］	Xu Linhong,Lin Hongfei,Pan Yu,et al.Constructing the affective lexicon ontology［J］.Journal of the China Society for Scientific and Technical Information,2008,27(2):180185.(in Chinese)
［7］	CLIWC ［EB/OL］.［20150220］.http://cliwc.blogbus.com/.
［8］	Russell J A.A circumplex model of affect［J］.Journal of Personality and Social Psychology,1980,39:11611178.
［9］	Read J. Recognising affect in text using pointwisemutual information［D］.UK:University of Sussex, 2004:141.
［10］	Mohammad S M,Turney P D.Emotions evoked by common words and phrases:Using mechanical turk to create an emotion lexicon［C］∥Proc of the NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text,2010:2634.
［11］	Strapparava C,Valitutti A.WordNetaffect:An affective extension of WordNet［C］∥Proc of LREC 2004,2004:11031106.
［12］	Xu J, Xu Reifeng, Zhang Yanzhen,et al.Chinese emotion lexicon developing via multilingual lexical resources integration［C］∥Proc of CICLing,2013:174182.
［13］	Kamps J,Marx M,Mokken R J,et al.Using WordNet to measure semantic orientation of adjectives［C］∥Proc of LREC04, the 4th International Conference on Language Resources and Evaluation,2004:11151118.
［14］	Zhu Yanlan,Min Jin,Zhou Yaqian,et al.Semantic orientation computing based on HowNet［J］.Journal of Chinese Information Processing,2006,20(1):1420.(in Chinese)
［15］	Lu Bin,Wan Xiaojun,Yang Jianwu,et al.Using Tongyici Cilin to compute word semantic polarity［C］∥Proc of ICCC2007 Chinese Information Processing Society of China,2007:410417.(in Chinese)
［16］	Turney P D,Littman M I.Measuring praise and criticism:Inference of semantic orientation from association［J］.ACM Transactions on Information Systems,2003,21(4):315346.
［17］	LIWC［EB/OL］.［20150220］.http://www.liwc.net/.
［18］	Golder S A,Macy M W.Diurnal and seasonal mood vary with work,sleep,and day length across diverse cultures［J］.Science,2011,333(6051):18781881.
［19］	NLPIR［EB/OL］.［20150220］.http://ictclas.nlpir.org/.
［20］	Huang Tinghao,Yu Hocheng,Chen Hsinhsi.Modeling pollyanna phenomena in Chinese sentiment analysis［C］∥Proc of COLING (Demos),2012:231238.
［21］	Xu Ge,Meng Xinfan,Wang Houfeng.Build Chinese emotion lexicons using a graphbased algorithm and multiple resources［C］∥Proc of the 23rd International Conference on Computational Linguistics,2010:12091217.
［22］	Liu Ning,Ren Fuji,Sun Xiao,et al.Microblogging hot events emotion analysis based on RenCECps［C］∥Proc of 2013 IEEE/SICE International Symposium on System Integration(SII),2013:233238.
	附中文参考文献：
［6］	徐琳宏,林鸿飞,潘宇,等.情感词汇本体的构造［J］.情报学报,2008,27(2):180185.
［14］	朱嫣岚,闵锦,周雅倩,等.基于HowNet的词汇语义倾向计算［J］.中文信息学报,2006,20(1):1420.
［15］	路斌,万小军,杨建武,等.基于同义词词林的词汇褒贬计算［C］∥第七届中文信息处理国际会议论文集,2007:410417.