基于双重注意力模型的微博情感倾向性分析

doi:10.13232/j.cnki.jnju.2020.02.009

南京大学学报(自然科学版) ›› 2020, Vol. 56 ›› Issue (2): 236–243.doi: 10.13232/j.cnki.jnju.2020.02.009

基于双重注意力模型的微博情感倾向性分析

罗春春,郝晓燕()

太原理工大学信息与计算机学院，晋中，030600

收稿日期:2020-01-13 出版日期:2020-03-30 发布日期:2020-04-02
通讯作者: 郝晓燕 E-mail:haoxiaoyan@tyut.edu.cn
基金资助:
教育部人文社会科学研究规划基金(17YJA740031);山西省自然科学基金(201801D121137)

Microblog sentiment orientation analysis based on double attention model

Chunchun Luo,Xiaoyan Hao()

School of Information and Computer, Taiyuan University of Technology, Jinzhong, 030600, China

Received:2020-01-13 Online:2020-03-30 Published:2020-04-02
Contact: Xiaoyan Hao E-mail:haoxiaoyan@tyut.edu.cn

摘要/Abstract

摘要：

在现有的微博情感倾向性分析任务中，微博标签往往被视为噪声信息，在数据预处理阶段就被剔除.但微博标签蕴含着微博内容的关键信息，所以标签的剔除对于微博的情感倾向性分析是不利的.针对该问题，充分考虑微博的文本特点，提出一种基于双重注意力的情感分析模型.采用Bi?LSTM （Bi?directional Long Short?Term Memory）分别构建微博文本和微博标签的语义表示，采用双重注意力机制同时对微博的正文层和微博的标签层进行语义编码，提取出文本中的关键信息.最后，基于所构建的语义表示训练情感分类模型.实验结果表明，该模型在微博情感倾向性分析上取得了较好的效果.

关键词: 双向长短期记忆网络, 双重注意力模型, 情感倾向性分析, 新浪微博, 微博标签

Abstract:

In the existing microblog emotion orientation classification task，the microblog tag is often regarded as noise information and eliminated in the data preprocessing stage. But the microblog tag contains key information and the elimination of tags is not good for microblog emotional orientation classification. Aiming at this problem，this paper fully considers the characteristics of microblog text and proposes a sentiment analysis model based on double attention model. The Bi?LSTM

（Bi?directional Long Short?Term Memory） is used to construct the semantic representation of the microblog text and the microblog tag. The double attention mechanism is used to semantically encode the microblog text and the microblog tag and to extract the key information in the text. Finally，the sentiment classification model is trained based on the constructed semantic representation. Experimental results show that the model achieved good results in the microblog emotional sentiment analysis.

Key words: Bi?directional Long Short?Term Memory, double attention model, sentiment orientation analysis, Sina microblog, microblog tag

中图分类号:

TP391.1

罗春春,郝晓燕. 基于双重注意力模型的微博情感倾向性分析[J]. 南京大学学报(自然科学版), 2020, 56(2): 236–243.

Chunchun Luo,Xiaoyan Hao. Microblog sentiment orientation analysis based on double attention model[J]. Journal of Nanjing University(Natural Sciences), 2020, 56(2): 236–243.

图/表 8

表1

图1

图2

表2

表3

表4

表5

图3

参考文献 23

1	黄佳峰，薛云，卢昕等. 面向中文网络评论情感分类的集成学习框架. 中文信息学报，2018，32(9)：113-122.
	Huang J F，Xue Y，Lu X， et al.An ensemble learning framework for sentiment classification of chinese online reviews. Journal of Chinese Information Processing，2018，32(9)：113-122.
2	王盛玉，曾碧卿，商齐等. 基于词注意力卷积神经网络模型的情感分析研究. 中文信息学报，2018，32(9)：123-131.
	Wang S Y，Zeng B Q，Shang Q， et al. Word attention?based convolutional neural networks for sentiment analysis. Journal of Chinese Information Processing，2018，32(9)：123-131.
3	孙建旺，吕学强，张雷瀚. 基于词典与机器学习的中文微博情感分析研究. 计算机应用与软件，2014，31(7)：177-181.
	Sun J W，Lu X Q，Zhang L H.On sentiment analysis of Chinese microblogging based on lexicon and machine learning. Computer Applications and Software，2014，31(7)：177-181.
4	Pang B，Lee L，Vaithyanathan S. Thumbs up：sentiment classification using machine learning techniques∥Proceedings of the ACL?02 Conference on Empirical Methods in Natural Language Processing. Stroudsburg，PA，USA：Association for Computational Linguistics，2002：79-86.
5	Kolajo T，Daramola O，Adebiyi A. Sentiment analysis on Naija?tweets∥Proceedings of the 57^th Annual Meeting of the Association for Computational Linguistics. Florence，Italy：ACM Press，2019：338-343.
6	周文，欧阳纯萍，阳小华等. 一种基于情感依存元组的简单句情感判别方法. 中文信息学报，2017，31(3)：177-183.
	Zhou W，Ouyang C P，Yang X H， et al. A simple?sentence sentiment classification method based onemotional dependency tuples. Journal of Chinese Information Processing，2017，31(3)：177-183.
7	王灿伟. 基于主题提取的海量微博情感分析. 南京大学学报(自然科学)，2017，53(3)：549-556.
	Wang C W.Sentimental analysis of massive micro?blog based on topic extraction. Journal of Nanjing University (Natural Science)，2017，53(3)：549-556.
8	汪鹏，赵学礼，李娜娜等. 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)，2017，53(6)：1133-1140.
	Wang P，Zhao X L，Li N N， et al. Research of cross?language sentiment classification based on structural correspondence learning. Journal of Nanjing University (Natural Science)，2017，53(6)：1133-1140.
9	杜慧，徐学可，伍大勇等. 基于情感词向量的微博情感分类. 中文信息学报，2017，31(3)：170-176.
	Du H，Xu X K，Wu D Y， et al. A sentiment classification method based on sentiment?specific word embedding. Journal of Chinese Information Processing，2017，31(3)：170-176.
10	任巨伟，杨亮，吴晓芳等. 基于情感常识的微博事件公众情感趋势预测. 中文信息学报，2017，31(2)：169-178. Ren J W，Yang L，Wu X F， et al. Public sentiment trend prediction of microblog events based on affective commonsense knowledge. Journal of Chinese Information Processing，2017，31(2)：169-178.
11	Thongtan T，Phienthrakul T. Sentiment classifi?cation using document embeddings trained with cosine similarity∥Proceedings of the 57^th Annual Meeting of the Association for Computational Linguistics：Student Research. Florence，Italy：ACM Press，2019：407-414.
12	Kim Y. Convolutional neural networks for sentence classification∥Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. Doha，Qatar：The Association for Computational Linguistics，2014.
13	Tang D，Qin B，Liu T，et al. Document modeling with gated recurrent neural network for sentiment classification∥Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Lisbon，Portugal：The Association for Computational Linguistics，2015：1422-1432.
14	Lai S W，Xu L H，Liu K，et al. Recurrent convolutional neural networks for text classification∥Proceedings of the 29^th AAAI Conference on Artificial Intelligence. Austin,TX,USA:AAAI，2015：2267-2273.
15	Zhou C T，Sun C，Liu Z，et al. A C?LSTM neural network for text classification. Computer Science，2015，1(4)：39-44.
16	Yang Z C，Yang D Y，Dyer C，et al. Hierarchical attention networks for document classification∥Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics：Human Language Technologies. San Diego,CA,USA:ACL，2016：1480-1489.
17	Bao L X,Lambert P,Badia T. Attention and lexicon regularized LSTM for aspect based sentiment analysis∥Proceedings of the 57^th Annual Meeting of the Association for Computational Linguistics. Florence，Italy：ACM Press，2019：253-259.
18	程艳，叶子铭，王明文等. 融合卷积神经网络与层次化注意力网络的中文文本情感倾向性分析. 中文信息学报，2019，33(1)：133-142.
	Cheng Y，Ye Z M，Wang M W， et al. Chinese text sentiment orientation analysis based on convolution neural network and hierarchical attention network. Chinese Journal of Information Science，2019，33(1)：133-142.
19	曾锋，曾碧卿，韩旭丽等. 基于双层注意力循环神经网络的方面级情感分析. 中文信息学报，2019，33(6)：108-115.
	Zeng F，Zeng B Q，Han X L， et al. Double attention neural network for aspect?based sentiment analysis. Journal of Chinese Information Processing，2019，33(6)：108-115.
20	Hochreiter S，Schmidhuber J. Long short?term memory. Neural Computation，1997，9(8)：1735-1780.
21	Tang D，Qin B，Feng X，et al. Effective LSTMs for target?dependent sentiment classification∥Proceedings of the 26^th International Conference on Computational Linguistics. Osaka，Japan：The Association for Computational Linguistics，2016：3298-3307.
22	Wang Y Q，Huang M，Zhu X，et al. Attention?based LSTM for aspect?level sentiment classification∥Proceedings of Conference on Empirical Methods in Natural Language Processing. Austin，TX,USA：The Association for Computational Linguistics，2016：606-615.
23	Wang W C，Feng S，Gao W，et al. Personalized microblog sentiment classification via adversarial cross?lingual multi?task learning∥Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Brussels，Belgium：The Association for Computational Linguistics，2018：338-348.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

数据集	正向	负向	总计
训练集	4561	3962	8523
测试集	1202	929	2131
总计	5763	4891	10654

参数名	参数值
词向量维度	200
批处理个数	32
学习率	0.001
dropout	0.2
LSTM输出维度	200

模型	Accuracy（%）
CNN	82.56
LSTM	83.09
LSTM+ATT	83.68
Bi?LSTM	83.83
SAM	84.06
Bi?LSTM+ATT	84.17
ACNN	82.87
H?ATT	84.40
DAM(本文模型)	85.39

基于双重注意力模型的微博情感倾向性分析

Microblog sentiment orientation analysis based on double attention model

RichHTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 23

相关文章 1

Metrics

本文评价

推荐阅读 10