基于结构对应学习的跨语言情感分类研究

南京大学学报(自然科学版) ›› 2017, Vol. 53 ›› Issue (6): 1133–.

基于结构对应学习的跨语言情感分类研究

汪　鹏1，2，赵学礼1，2，李娜娜1，2，董永峰1，2*

出版日期:2017-11-27 发布日期:2017-11-27
作者简介:1.河北工业大学计算机科学与软件学院，天津，300401；
2．河北省大数据计算重点实验室(河北工业大学)，天津，300401
基金资助:
基金项目：天津市自然科学基金(14JCYBJC18500)，天津市应用基础与前沿技术研究计划(13JCQNJC00200)
收稿日期：2017－08－31
*通讯联系人，E－mail：dongyongfeng@scse.hebut.edu.cn

Research of cross－language sentiment classification based on structural correspondence learning

Wang Peng1，2，Zhao Xueli1，2，Li Nana1，2，Dong Yongfeng1，2*

Online:2017-11-27 Published:2017-11-27
About author:1.School of Computer Science & Engineering，Hebei University of Technology，Tianjin，300401，China；
2．Hebei Province Key Laboratory of Big Data Calculation(Hebei University of Technology)，Tianjin，300401，China

摘要/Abstract

摘要： 情感分类的主要目的是预测用户在互联网中发布情绪数据的极性(积极的或者消极的)，各种语言的情感分析已经成为诸多应用的研究热点，然而由于不同语言的情感资源在质量和数量上的不平衡，通常使用源语言来改善目标语言的跨语言情感分类方法，来提高目标语言情感分类的准确性．传统的跨语言情感分类主要是通过机器翻译将目标语言映射到源语言中，但是分类的准确性严重受到机器翻译质量的影响．通过对跨领域文本分类的结构学习算法(SCL)的讨论和拉普拉斯映射对两种语言之间词对的影响，对跨语言结构对应学习算法(CLSCL)的改进，进而提出M－CLSCL算法，借助选出来的轴心词对来进行目标语言的情感分类，通过M－CLSCL方法与前述相关方法的实验结果进行比较，可以发现M－CLSCL提高了情感分类的准确性．

Abstract: The main purpose of sentiment classification is to predict the sentiment polarity(positive or negative)of the data that the user publishes the emotional data on the Internet.The sentiment analysis of various languages has become a hotspot of many applications，but because of the different quality of sentiment resources and quantity on the imbalance，the researchers usually use the source language to improve the target language cross－language sentiment classification method to improve the accuracy of the target language sentiment classification.The traditional cross－language sentiment classification mainly maps the target language to the source language through machine translation，but the accuracy of the classification is seriously influenced by the quality of the machine translation.This paper discusses the structure of the cross－language(CLSCL)，and then put forward the M－CLSCL algorithm，by using the selected axis of the word to the target language of the language of the target language sentiment classification.Comparing the experimental results of M－CLSCL method and the relevant methods，we find that M－CLSCL has improved the accuracy of emotional classification.

汪　鹏1，2，赵学礼1，2，李娜娜1，2，董永峰1，2*. 基于结构对应学习的跨语言情感分类研究[J]. 南京大学学报(自然科学版), 2017, 53(6): 1133–.

Wang Peng1，2，Zhao Xueli1，2，Li Nana1，2，Dong Yongfeng1，2*. Research of cross－language sentiment classification based on structural correspondence learning[J]. Journal of Nanjing University(Natural Sciences), 2017, 53(6): 1133–.

参考文献

[1]　Lin Z，Jin X L，Xu X K，et al.A cross－lingual joint aspect/sentiment model for sentiment analysis.In：Proceedings of the 23th ACM International Conference on Information and Knowledge Management.Shanghai，China：ACM，2014：1089－1098.
[2] Wei B，Pal C．Cross lingual adaptation：An experiment on sentiment classifications.In：Proceedings of the ACL 2010 Conference Short Papers.Uppsala，Sweden：ACM，2010：258－262.
[3] 赵妍妍，秦　兵，刘　挺．文本情感分析．软件学报，2010，21(8)：1834－1848.(Zhao Y Y，Qin B，Liu T．Sentiment analysis.Journal of Software，2010，21(8)：1834－1848.)
[4] Kennedy A，Inkpen D．Sentiment classification of movie reviews using contextual valence shifters.Computational Intelligence，2006，22(2)：110－125.　
[5] Gui L，Xu R F，Xu J，et al.A mixed model for cross lingual opinion analysis.In：Zhou G D，Li J Z，Zhao D Y，et al.Natural Language Processing and Chinese Computing.Springer Berlin Heidelberg，2013：93－104.
[6] Li S S，Wang R，Liu H H，et al.Active learning for cross－lingual sentiment classification.In：Zhou G D，Li J Z，Zhao D Y，et al.Natural Language Processing and Chinese Computing.Springer Berlin Heidelberg，2013：236－246.
[7] Chen Q，He Y X，Liu X L，et al.Cross－language sentiment analysis based on parser.Acta Scientiarum Naturalium Universitatis Pekinensis，2014，50(1)：55－60.
[8] Prettenhofer P，Stein B．Cross－lingual adaptation using structural correspondence learning.ACM Transactions on Intelligent Systems and Technology，2011，3(1)：13.
[9] Blitzer J，McDonald R，Pereira F．Domain adaptation with structural correspondence learning.In：Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing.Sydney，Australia：ACM，2006：120－128.　
[10] Pan S J，Ni X，Sun J T，et al.Cross－domain sentiment classification via spectral feature alignment.In：Proceedings of the 19th International Conference on World Wide Web.Raleigh，NC，USA：ACM，2010：751－760.
[11] Sarath C A P，Lauly S，Larochelle H，et al.An autoencoder approach to learning bilingual word representations.In：Proceedings of the 28th Neural Information Processing Systems.Montreal，Canada：NIPS Press，2014，3：1853－1861.
[12] Zhou H W，Chen L，Shi F L，et al.Learning bilingual sentiment word embeddings for cross－language sentiment classification.In：Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing.Beijing，China：ACL，2015：430－440.
[13] Tang X W，Wan X J．Learning bilingual embedding model for cross－language sentiment classification.In：Proceedings of the IEEE/WIC/ACM International Joint Web Intelligence(WI)and Intelligent Agent Technologies(IAT)．Warsaw，Poland：IEEE，2014：134－141.
[14] Zhou X J，Wan X J，Xiao J G．Cross－lingual sentiment classification with bilingual document representation learning.In：Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics.Berlin，Germany：ACL，2016：1403－1412.
[15] 范小丽，刘晓霞．文本分类中互信息特征选择方法的研究．计算机工程与应用，2010，46(34)：123－125.(Fan X L，Liu X X．Study on mutual information－based feature selection in text categorization.Computer Engineering and Applications，2010，46(34)：123－125.)
[16] Web Technology and Information Systems.http：//www.webis.de/research/corpora/corpus－webis－cls－10/cls－acl10－processed.tar.gz，2010.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed