基于多特征融合与双向RNN的细粒度意见分析
|
摘要
文本细粒度意见分析主要有属性抽取和基于属性的情感分类2个任务,现有方法完成上述任务采用条件随机场(CRF)训练属性抽取模型,并运用循环神经网络(RNN)训练基于属性的情感分类模型。但同时完成2个任务则无法找到属性和情感倾向的对应关系。针对该问题,提出利用双向RNN构建基于序列标注的细粒度意见分析模型。通过融合文本的词向量、词性和依存关系等语言学特征,学习文本的修饰和语义信息,并设计一个时间序列标注模型,同时抽取属性实体判断文本的情感极性。在真实数据集上的实验结果表明,与CRF、TD-LSTM、AELSTM等模型相比,该模型情感分类效果提升明显。
Text fine-grained opinion analysis mainly includes attribute extraction and attribute-based sentiment classification. The existing methods accomplish the above tasks by adopting Conditional Random Field( CRF) training attribute extraction model and using a Recurrent Neural Network( RNN) to train attribute-based emotion classification model. However,the completion of two tasks at the same time can not find the corresponding relationship between attributes and emotional tendencies. To solve this problem,a two-dimensional RNN is proposed to build a fine-grained opinion analysis model based on sequence annotation. By merging the linguistic features of the text, such as word vectors,part of speech and dependence,it learns the text's modification and semantic information,designs a time series annotation model, and extracts attribute entities to determine the sentiment polarity of the text. Experimental results on real datasets show that compared with CRF,TD-LSTM,AE-LSTM and other models,the emotional classification effect of this model obviously improves.
Text fine-grained opinion analysis mainly includes attribute extraction and attribute-based sentiment classification. The existing methods accomplish the above tasks by adopting Conditional Random Field( CRF) training attribute extraction model and using a Recurrent Neural Network( RNN) to train attribute-based emotion classification model. However,the completion of two tasks at the same time can not find the corresponding relationship between attributes and emotional tendencies. To solve this problem,a two-dimensional RNN is proposed to build a fine-grained opinion analysis model based on sequence annotation. By merging the linguistic features of the text, such as word vectors,part of speech and dependence,it learns the text's modification and semantic information,designs a time series annotation model, and extracts attribute entities to determine the sentiment polarity of the text. Experimental results on real datasets show that compared with CRF,TD-LSTM,AE-LSTM and other models,the emotional classification effect of this model obviously improves.
引文
[1]章剑锋,张奇,吴立德,等.中文观点挖掘中的主观性关系抽取[J].中文信息学报,2008,22(2):55-59.
[2]赵妍妍,秦兵,刘挺.文本情感分析[J].软件学报,2010,21(8):1834-1848.
[3]姚天昉,程希文,徐飞玉,等.文本意见挖掘综述[J].中文信息学报,2008,22(3):71-80.
[4]厉小军,戴霖,施寒潇,等.文本倾向性分析综述[J].浙江大学学报(工学版),2011,45(7):1167-1174.
[5]HAO Z,CHEN J,CAI R,et al.Chinese sentiment classification based on the sentiment drop point[C]//Proceedings of International Conference on Intelligent Computing.Berlin,Germany:Springer,2013:55-60.
[6]LIU B.Sentiment analysis and opinion mining[J].Synthesis Lectures on Human Language Technologies,2012,5(1):155-167.
[7]HU M,LIU B.Mining and summarizing customer reviews[C]//Proceedings of the 20th ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining.New York,USA:ACM Press,2004:168-177.
[8]POPESCH A M,ETZIONI O.Extracting product features and opinions from reviews[M].Berlin,Germany:Springer,2007:9-28.
[9]JAKOB N,GUREYYCH I.Extracting opinion targets in a single-and cross-domain setting with conditional random fields[C]//Proceedings of 2010 Conference on Empirical Methods in Natural Language Processing,[S.l.]:Association for Computational Linguistics,2010:1035-1045.
[10]HAO Z,WANG H,CAI R,et al.Product named entity recognition for Chinese query questions based on a skipchain CRF model[J].Neural Computing and Applications,2013,23(2):371-379.
[11]ZHUANG L,JING F,ZHU X Y.Movie review mining and summarization[C]//Proceedings of the 15th ACMInternational Conference on Information and Knowledge Management.New York,USA:ACM Press,2006:43-50.
[12]KOBAYASHI N,INUI K,MATSUMOTO Y.Extracting aspect-evaluation and aspect-of relations in opinion mining[C]//Proceedings of EMNLP-CoNLL’07.Washington D.C.,USA:IEEE Press,2007:1065-1074.
[13]杨凤芹,宋美佳,孙铁利,等.面向中文产品评论的完整评价对象抽取方法[J].计算机工程,2017,43(6):169-176.
[14]BENGIO Y,DUCHARME R,VINCENT P,et al.Aneural probabilistic language model[J].Journal of Machine Learning Research,2003,3(1):1137-1155.
[15]MIKALOV T,KARAFIAT M,BURGET L,et al.Recurrent neural network based language model[C]//Proceedings of Conference on International Speech Communication Association.Chiba,Japan:[s.n.],2010:1045-1048.
[16]MIKOLOV T,SUTSKEVER I,CHEN K,et al.Distributed representations of words and phrases and their compositionality[C]//Proceedings of IEEEInternational Conference on Neural Information Processing Systems.Washington D.C.,USA:IEEEPress,2013:3111-3119.
[17]IRSOY O,CARDIE C.Opinion mining with deep recurrent neural networks[C]//Proceedings of IEEEConference on Empirical Methods in Natural Language Processing.Washington D.C.,USA:IEEE Press,2014:720-728.(下转第页204
[18]LIU P,JOTY S,MENG H.Fine-grained opinion mining with recurrent neural networks and word embeddings[C]//Proceedings of ACM Conference on Empirical Methods in Natural Language Processing.New York,USA:ACM Press,2015:1433-1443.
[19]WANG B,LIU M.Deep learning for aspect-based sentiment analysis[EB/OL].[2017-05-21].http://web.stanford.edu/class.
[20]TANG D,QIN B,FENG X,et al.Effective LSTMs for target-dependent sentiment classification[EB/OL].[2017-05-21].http://www.aclweb.org/.
[21]TANG D,QIN B,LIU T.Aspect level sentiment classification with deep memory network[EB/OL].[2017-05-21].http://wing.comp.nus.edu.sg/.
[22]SUKHBAATAR S,WESTON J,FERGUS R.End-toend memory networks[J].Neural Information Processing Systems,2015(1):2440-2448.
[23]WANG Y,HUANG M,ZHU X,et al.Attention-based LSTM for aspect-level sentiment classification[C]//Proceedings of IEEE EMNLP’06.Washington D.C.,USA:IEEE Press,2016:606-615.
[24]ZHANG Y,ZHANG H,ZHANG M,et al.Do users rate or review?:boost phrase-level sentiment labeling with review-level sentiment classification[C]//Proceedings of the 37th International ACM SIGIR Conference on Research&Development in Information Retrieval.New York,USA:ACM Press,2014:1027-1030.编辑索书志
[2]赵妍妍,秦兵,刘挺.文本情感分析[J].软件学报,2010,21(8):1834-1848.
[3]姚天昉,程希文,徐飞玉,等.文本意见挖掘综述[J].中文信息学报,2008,22(3):71-80.
[4]厉小军,戴霖,施寒潇,等.文本倾向性分析综述[J].浙江大学学报(工学版),2011,45(7):1167-1174.
[5]HAO Z,CHEN J,CAI R,et al.Chinese sentiment classification based on the sentiment drop point[C]//Proceedings of International Conference on Intelligent Computing.Berlin,Germany:Springer,2013:55-60.
[6]LIU B.Sentiment analysis and opinion mining[J].Synthesis Lectures on Human Language Technologies,2012,5(1):155-167.
[7]HU M,LIU B.Mining and summarizing customer reviews[C]//Proceedings of the 20th ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining.New York,USA:ACM Press,2004:168-177.
[8]POPESCH A M,ETZIONI O.Extracting product features and opinions from reviews[M].Berlin,Germany:Springer,2007:9-28.
[9]JAKOB N,GUREYYCH I.Extracting opinion targets in a single-and cross-domain setting with conditional random fields[C]//Proceedings of 2010 Conference on Empirical Methods in Natural Language Processing,[S.l.]:Association for Computational Linguistics,2010:1035-1045.
[10]HAO Z,WANG H,CAI R,et al.Product named entity recognition for Chinese query questions based on a skipchain CRF model[J].Neural Computing and Applications,2013,23(2):371-379.
[11]ZHUANG L,JING F,ZHU X Y.Movie review mining and summarization[C]//Proceedings of the 15th ACMInternational Conference on Information and Knowledge Management.New York,USA:ACM Press,2006:43-50.
[12]KOBAYASHI N,INUI K,MATSUMOTO Y.Extracting aspect-evaluation and aspect-of relations in opinion mining[C]//Proceedings of EMNLP-CoNLL’07.Washington D.C.,USA:IEEE Press,2007:1065-1074.
[13]杨凤芹,宋美佳,孙铁利,等.面向中文产品评论的完整评价对象抽取方法[J].计算机工程,2017,43(6):169-176.
[14]BENGIO Y,DUCHARME R,VINCENT P,et al.Aneural probabilistic language model[J].Journal of Machine Learning Research,2003,3(1):1137-1155.
[15]MIKALOV T,KARAFIAT M,BURGET L,et al.Recurrent neural network based language model[C]//Proceedings of Conference on International Speech Communication Association.Chiba,Japan:[s.n.],2010:1045-1048.
[16]MIKOLOV T,SUTSKEVER I,CHEN K,et al.Distributed representations of words and phrases and their compositionality[C]//Proceedings of IEEEInternational Conference on Neural Information Processing Systems.Washington D.C.,USA:IEEEPress,2013:3111-3119.
[17]IRSOY O,CARDIE C.Opinion mining with deep recurrent neural networks[C]//Proceedings of IEEEConference on Empirical Methods in Natural Language Processing.Washington D.C.,USA:IEEE Press,2014:720-728.(下转第页204
[18]LIU P,JOTY S,MENG H.Fine-grained opinion mining with recurrent neural networks and word embeddings[C]//Proceedings of ACM Conference on Empirical Methods in Natural Language Processing.New York,USA:ACM Press,2015:1433-1443.
[19]WANG B,LIU M.Deep learning for aspect-based sentiment analysis[EB/OL].[2017-05-21].http://web.stanford.edu/class.
[20]TANG D,QIN B,FENG X,et al.Effective LSTMs for target-dependent sentiment classification[EB/OL].[2017-05-21].http://www.aclweb.org/.
[21]TANG D,QIN B,LIU T.Aspect level sentiment classification with deep memory network[EB/OL].[2017-05-21].http://wing.comp.nus.edu.sg/.
[22]SUKHBAATAR S,WESTON J,FERGUS R.End-toend memory networks[J].Neural Information Processing Systems,2015(1):2440-2448.
[23]WANG Y,HUANG M,ZHU X,et al.Attention-based LSTM for aspect-level sentiment classification[C]//Proceedings of IEEE EMNLP’06.Washington D.C.,USA:IEEE Press,2016:606-615.
[24]ZHANG Y,ZHANG H,ZHANG M,et al.Do users rate or review?:boost phrase-level sentiment labeling with review-level sentiment classification[C]//Proceedings of the 37th International ACM SIGIR Conference on Research&Development in Information Retrieval.New York,USA:ACM Press,2014:1027-1030.编辑索书志