计算机科学 ›› 2023, Vol. 50 ›› Issue (5): 238-247.doi: 10.11896/jsjkx.220400256

• 人工智能 • 上一篇    下一篇

基于多事件语义增强的情感分析

张雪, 赵晖   

  1. 新疆大学信息科学与工程学院 乌鲁木齐 830046
    新疆维吾尔自治区信号检测与处理重点实验室 乌鲁木齐 830046
    新疆维吾尔自治区多语种信息技术重点实验室 乌鲁木齐 830046
  • 收稿日期:2022-04-26 修回日期:2022-09-05 出版日期:2023-05-15 发布日期:2023-05-06
  • 通讯作者: 赵晖(zhmerry@126.com)
  • 作者简介:(zhangx_0803@foxmail.com)
  • 基金资助:
    国家自然科学基金(62166041)

Sentiment Analysis Based on Multi-event Semantic Enhancement

ZHANG Xue, ZHAO Hui   

  1. College of Information Science and Engineering,XinjiangUniversity,Urumqi 830046,China
    Key Laboratory of Signal Detection and Processing,Xinjiang Uygur Autonomous Region,Urumqi 830046,China
    Key Laboratory of Multilingual Information Technology, Xinjiang Uygur Autonomous Region,Urumqi 830046,China
  • Received:2022-04-26 Revised:2022-09-05 Online:2023-05-15 Published:2023-05-06
  • About author:ZHANG Xue,born in 1995,postgra-duate.Her main research interests include natural language processing and implicit sentiment analysis.
    ZHAO Hui,born in 1972 Ph.D,professor,Ph.D supervisor,is a member of China Computer Federation.Her main research interests include artificial intelligence,affective computing,speech and digital image processing.
  • Supported by:
    National Natural Science Foundation of China(62166041).

摘要: 隐式情感分析是检测不包含明显情感词的句子的情感。文中集中于以事件为中心的情感分析,该任务是通过句子中描述的事件推断其情感极性。在以事件为中心的情感分析中,现有方法要么将文本中名词短语看作事件,要么采用复杂的模型建模事件,未能充分建模事件信息,并且没有考虑到包含多个事件的情况。为解决以上问题,提出将事件表示为事件三元组〈主语,谓语,宾语〉的形式,基于这种事件表示,进一步提出基于事件增强语义的情感分析模型MEA来检测文本的情感。文中利用句法信息捕获事件三元组的关系,根据每个事件对句子贡献程度的不同,采用注意力机制建模事件之间的关系。与此同时,采用双向长短时记忆网络建模句子的上下文信息,并采用多级性正交注意力机制捕获不同极性下注意力权重的差异,这可以作为显著的判别特征。最后,依据事件特征和句子特征的重要程度为其分配不同的权重比例,并将它们融合得到最终的句子表示。此外,文中还提出一个用于事件增强情感分析的数据集MEDS,其中每条句子都标有事件三元组表示和情感极性标签。研究表明,在自建的数据集中,所提模型优于现有的基线模型。

关键词: 事件型情感分析, 表示学习, 情感分析, 图卷积神经网络, 注意力机制

Abstract: Implicit sentiment analysis is to detect the sentiment of sentences that do not contain obvious sentiment words.This paper focuses on event-centric sentiment analysis,which is to infer sentiment polarity from the events described in the sentence.In event-centric sentiment analysis,existing methods either treat noun phrases in the text as events,or adopt complex models to model events,fail to model event information sufficiently,and fail to consider events that contain multiple events.In order to solve the above problems,it is proposed to represent events in the form of event triples〈subject,predicate,object〉.Based on this event representation,an event-enhanced semantic-based sentiment analysis model(MEA) is further proposed to detect the sentiment of texts.In this paper,syntactic information is used to capture the relationship of event triples,and attention mechanism is used to model the relationship between events according to the contribution of each event to the sentence.At the same time,a bidirec-tional long-short-term memory network(Bi-LSTM) is used to model the contextual information of sentences,and a multi-level orthogonal attention mechanism is used to capture the difference of attention weights under different polarities,which can be used as a significant discriminative feature.Finally,according to the importance of event features and sentence features,they are assigned different weight ratios,and they are fused to obtain the final sentence representation.Furthermore,this paper proposes a dataset for event-enhanced sentiment analysis(MEDS),where each sentence is labeled with event triplet representations and sentiment polarity labels.Research shows that the proposed model outperforms existing baseline models in self-built datasets.

Key words: Event-based sentiment analysis, Representation learning, Sentiment analysis, Graph convolutional neural network, Attention mechanism

中图分类号: 

  • TP391
[1]LIU B.Sentiment analysis:Mining opinions,sentiments,andemotions[M].Cambridgeshire:Cambridge University Press,2020.
[2]CAMBRIA E,PORIA S,GELBUKH A,et al.Sentiment analysis is a big suitcase[J].Intelligent Systems,2018,32(6):74-80.
[3]LIAO J,WANG S,LI D.Identification of fact-implied implicitsentiment based on multi-level semantic fused representation[J].Knowledge-Based Systems,2019,165:197-207.
[4]GREENE S,RESNIK P.More than words:Syntactic packaging and implicit sentiment[C]//Proceedings of Human Language Technologies:The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics.2009:503-511.
[5]ZHOU D,WANG J,ZHANG L,et al.Implicit Sentiment Ana-lysis with Event-centered Text Representation[C]//Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing.2021:6884-6893.
[6]WEBER N,BALASUBRAMANIAN N,CHAMBERS N.Event representations with tensor-based compositions[C]//Procee-dings of the I Conference on Artificial Intelligence.2018.
[7]ZHANG C,LI Q,SONG D.Aspect-based sentiment classification with aspect-specific graph convolutional networks[J].ar-Xiv:1909.03477,2019.
[8]SUN K,ZHANG R,MENSAH S,et al.Aspect-level sentiment analysis via convolution over dependency tree[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing(EMNLP-IJCNLP).2019:5679-5688.
[9]SHUTOVA E,TEUFEL S,KORHONEN A.Statistical metaphor processing[J].Computational Linguistics,2013,39(2):301-353.
[10]ZHANG L,LIU B.Identifying noun product features that imply opinions[C]//Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics:Human Language Technologies.2011:575-580.
[11]BALAHUR A,HERMIDA J M,MONTOYO A.Detecting implicit expressions of sentiment in text based on commonsense knowledge[C]//Proceedings of the 2nd Workshop on Computational Approaches to Subjectivity and Sentiment Analysis(WASSA 2.011).2011:53-60.
[12]DENG L J,WIEBE J.Joint prediction for entity/event-level sentiment analysis using probabilistic soft logic models[C]//Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing.2015:179-189.
[13]HOFMANN J,TROIANO E,SASSENBERG K,et al.Appraisal theories for emotion classification in text[J].arXiv:2003.14155,2020.
[14]GAONKAR R S.Modeling label semantics for predicting emotional reactions[D].Stony Brook:State University of New York at Stony Brook,2019.
[15]DING H,RILOFF E.Acquiring knowledge of affective events from blogs using label propagation[C]//Thirtieth I Conference on Artificial Intelligence.2016:2935-2942.
[16]DING H,RILOFF E.Weakly supervised induction of affectiveevents by optimizing semantic consistency[C]//Proceedings of the I Conference on Artificial Intelligence.2018.
[17]DING H,RILOFF E.Human needs categorization of affectiveevents using labeled and unlabeled data[C]//Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies,Volume 1(Long Papers).2018:1919-1929.
[18]YANG Y,ZHOU D,HE Y,et al.Interpretable relevant emotion ranking with event-driven attention[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing(EMNLP-IJCNLP).2019:177-187.
[19]ZHUANG Y,JIANG T,RILOFF E.Affective event classification with discourse-enhanced self-training[C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing(EMNLP).2020:5608-5617.
[20]ROSENTHAL S,FARRA N,NAKOV P.SemEval-2017 task 4:Sentiment analysis in Twitter[C]//Proceedings of the 11th international workshop on semantic evaluation(SemEval-2017).2017:502-518.
[21]GRISHMAN R,WESTBROOK D,MEYERS A.Nyu's English ace 2005 system description[C]//ACE 2005 Evaluation Workshop.2005.
[22]HE L,LEE K,LEWIS M,et al.Deep semantic role labeling:What works and what's next[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics(Volume 1:Long Papers).2017:473-483.
[23]PENNINGTON J,SOCHER R,MANNING C D.GloVe:Global vectors for word representation[C]//Proceedings of the 2014 Conference on Empirical Methods in Natural Language Proces-sing(EMNLP).Doha,2014:1532-1543.
[24]WEI J,LIAO J,YANG Z,et al.Bi-LSTM with multi-polarityorthogonal attention for implicit sentiment analysis[J].Neurocomputing,2020,383:165-173.
[25]SARA R,PRESLAV N,SVETLANA K,et al.SemEval-2015 Task 10:Sentiment Analysis in Twitter[C]//Proceedings of the 9th International Workshop on Semantic Evaluation(SemEval 2015).2015:451-463.
[26]CHEN D,MANNING C.A fast and accurate dependency parser using neural networks[C]//Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing(EMNLP).2014:740-750.
[27]KIM Y.Convolutional Neural Networks for Sentence Classification[J].arXiv:1408.5882,2014.
[28]LIU P F,QIU X P,HUANG X J.Recurrent Neural Network for Text Classification with Multi-Task Learning[C]//Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence.New York,USA:I Press,2014.2873-2879.
[29]ZHOU P,SHI W,TIAN J,et al.Attention-based bidirectional long short-term memory networks for relation classification[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics(volume 2:Short papers).2016:207-212.
[30]CHUNG J,GULCEHRE C,CHO K,et al.Empirical evaluation of gated recurrent neural networks on sequence modeling[J].arXiv:1412.3555,2014.
[31]GAO G,CHOI E,CHOI Y,et al.Neural metaphor detection in context[C]//Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing.2018.
[32]LIN Y,LI J P,XU K,et al.Sentiment Analysis with Multi-Head Attention-Based Bi-LSTM Model[J].Journal of Shanxi University,2020,43(1):1-7.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!