当前位置>主页 > 期刊在线 > 信息技术 >

信息技术23年3期

基于 BERT 预处理模型的网络舆情细粒度情感分析
徐子昂
(暨南大学深圳旅游学院,广东 深圳 518053)

摘  要:自互联网诞生以来,因其便利性、传播性和高自由度逐渐成为舆论的主要发酵地,也引起了网络空间中的舆论治理问题。在越来越多的社会事件中,舆论影响了整个事件的走向或者处理方式,甚至对事件中的相关人员产生影响,因此对舆情监控已然成为一个急需解决的问题。但舆情的负面性、矛盾性和复杂性也为监管增加了难度。为了推进舆情监控发展,研究使用基于 BERT 预处理模型的 E2E-ABSA,通过比较其他模型的表现来判断模型的可靠性,并与其他研究中使用的情感分析模型进行比较,并得出结论细粒度情感分析在评论携带多个主题且观点不一致的场景下具有明显优势。


关键词:BERT;E2E-ABSA;神经网络;深度学习;舆情监控;网络治理



DOI:10.19850/j.cnki.2096-4706.2023.03.003


中图分类号:TP391                                           文献标识码:A                                文章编号:2096-4706(2023)03-0014-06


Network Public Opinion Fine-Grained Emotion Analysis Based on BERT Preprocessing Model

XU Ziang

(Shenzhen Campus, Jinan University, Shenzhen 518053, China)

Abstract: Since the birth of the internet, because of its convenience, communication and high degree of freedom, it has gradually become the main fermentation place of public opinion, which has also caused problems of public opinion governance in cyberspace. In more and more social events, public opinion has affected the trend or handling method of the whole event, and even affected the relevant personnel in the event. Therefore, the monitoring of public opinion has become an urgent problem to be solved. However, the negative, contradictory and complex features of public opinion also increase the difficulty of supervision. In order to promote the development of public opinion monitoring, this paper researches and uses E2E-ABSA based on BERT preprocessing model. It judges the reliability of this model by comparing the performance with other models, and compares with the emotion analysis model used in other researches, and concludes that fine-grained emotion analysis in comments with multiple topics and scenario of inconsistent views has obvious advantages.

Keywords: BERT; E2E-ABSA; neural network; deep learning; public opinion monitoring; network governance


参考文献:

[1] 中国互联网络信息中心 .CNNIC 发布第 49 次《中国互联网络发展状况统计报告》 [J]. 新闻潮,2022(2):3.

[2] 彭知辉 . 论群体性事件与网络舆情 [J]. 公安理论与实践:上海公安高等专科学校学报,2008(1):46-50.

[3] XUE W,ZHOU W B,LI T,et al. MTNA:A Neural Multi-Task Model for Aspect Category Classification and Aspect Term Extraction on Restaurant Reviews [C]//the Eighth International Joint Conference on Natural Language Processing.Taipei:IJCNLP,2017,2:151-156.

[4] DEVLIN J,CHANG M W,LEE K,et al. BERT: Pre-Training of Deep Bidirectional Transformers for Language Understanding [J/OL].arXiv:1810.04805 [cs.CL].[2022-08-09].https:// arxiv.org/abs/1810.04805v1. 

[5] KAPIL P. A Novel Neural Network Architecture for SentimentOriented Aspect-Opinion Pair Extraction [D].Bangalore:Indian Institute  of Science,2021. 

[6] ZHANG W X,LI X,DENG Y,et al. A Survey on AspectBased Sentiment Analysis:Tasks,Methods,and Challenges [J/ OL].arXiv:2203.01054 [cs.CL].[2022-08-07].https://arxiv.org/ abs/2203.01054.

[7] FAN F F,FENG Y S,ZHAO D Y. Multi-Grained Attention Network for Aspect-Level Sentiment Classification [C]//the 2018 Conference on Empirical Methods in Natural Language Processing. Brussels:EMNLP,2018:3433-3442.

[8] 杨应全,刘子辉 . 我国突发事件网络舆情研究综述 [J]. 重庆科技学院学报:社会科学版,2022(2):40-48+71.

[9] 曲成义 . 网络突发事件的应急对策与治理环境的建设 [J].信息化建设,2005(10):16-17.

[10] 徐晓日 . 网络舆情事件的应急处理研究 [J]. 华北电力大学学报:社会科学版,2007(1):89-93.

[11] 史伟,薛广聪,何绍义 . 情感视角下的网络舆情研究综述 [J]. 图书情报知识,2022,39(1):105-118.

[12] 曾子明,万品玉 . 融合演化特征的公共安全事件微博情感分析 [J]. 情报科学,2018,36(12):3-8+51.

[13] 栗雨晴,礼欣,韩煦,等 . 基于双语词典的微博多类情感分析方法 [J]. 电子学报,2016,44(9):2068-2073.

[14] 孙雷 . 基于主题模型的微博评论细粒度情感分析研究[D]. 邯郸:河北工程大学,2021.

[15] SANG E F T K,VEENSTRA J. Representing Text Chunks [J/OL].arXiv:cs/9907006 [cs.CL].[2022-08-08].https://arxiv.org/abs/ cs/9907006.

[16] DEVLIN J,CHANG M W,LEE K,et al. BERT:Pretraining of Deep Bidirectional Transformers for Language Understanding [J/OL].arXiv:1810.04805 [cs.CL].[2022-08-09].https://arxiv.org/ abs/1810.04805v1.

[17] LI X,BING L D,ZHANG W X,et al. Exploiting BERT for End-to-End Aspect-Based Sentiment Analysis [J/OL].arXiv: 1910.00883 [cs.CL].[2022-08-06].https://arxiv.org/abs/1910.00883.

[18] CHO K,MERRIENBOER B V,GULCEHRE C,et al. Learning Phrase Representations Using RNN Encoder-Decoder for Statistical Machine Translation [J/OL].arXiv:1406.1078 [cs.CL].[2022-08-06].https://arxiv.org/abs/1406.1078.

[19] JOZEFOWICZ R,ZAREMBA W,SUTSKEVER I. An Empirical Exploration of Recurrent Network Architectures [C]//the 32nd International Conference on Machine Learning.Lille:ICML,2015,37:2342-2350.

[20] CHENG J P,DONG L,LAPATA M. Long Short-Term Memory-Networks for Machine Reading [J/OL].arXiv:1601.06733 [cs. CL].[2022-08-06].https://arxiv.org/abs/1601.06733v5.

[21] VASWANI A,SHAZEER N,PARMAR N,et al. Attention is all You Need [J/OL].arXiv:1706.03762 [cs.CL].[2022-08-08].https:// arxiv.org/abs/1706.03762.

[22] HUANG Z H,XU W,YU K. Bidirectional LSTM-CRF Models for Sequence Tagging [J/OL].arXiv:1508.01991 [cs.CL].[2022-08-09].https://arxiv.org/abs/1508.01991.

[23] LI X,BING L D,LI P J,et al. A Unified Model for Opinion Target Extraction and Target Sentiment Prediction [J/OL].arXiv: 1811.05082 [cs.CL].[2022-08-09].https://arxiv.org/abs/1811.05082.

[24] LAMPLE G,BALLESTEROS M,SUBRAMANIAN S, et al. Neural architectures for Named Entity Recognition [J/OL].arXiv: 1603.01360 [cs.CL].[2022-08-05].https://arxiv.org/abs/1603.01360.

[25] 谭翠萍 . 文本细粒度情感分析研究综述 [J]. 大学图书馆学报,2022,40(4):85-99+119.

[26] 张继东,张慧迪 . 融合注意力机制的多模态突发事件用户情感分析 [J]. 情报理论与实践,2022,45(11):170-177.

[27] 杨晓晨 . 基于深度学习的文本情感分类方法研究 [D]. 天津:天津理工大学,2022.

[28] LI X,BING L D,LI P J,et al. A Unified Model for Opinion Target Extraction and Target Sentiment Prediction [J/OL].arXiv: 1811.05082 [cs.CL].[2022-08-07].https://arxiv.org/abs/1811.05082.

[29] 华玮,吴思洋,俞超,吴婕洵,徐健 . 面向网络舆情事件的多层次情感分歧度分析方法 [J/OL]. 数据分析与知识发现,2022:1-23[2022-07-09].http://kns.cnki.net/kcms/ detail/10.1478.g2.20220613.1153.008.html.

[30] 王宇琪,郑安琪,郝川艳 . 基于细粒度情感分析的情感目标过程性评价机理与路径研究 [J]. 计算机时代,2022(7):94-97+100.

[31] 苏红旗,聂金慧,曾志 . 基于特征聚类的网络评论信息挖掘研究 [J]. 信息网络安全,2013(12):51-54.

[32] 赵宏,傅兆阳,王乐 . 基于特征融合的中文文本情感分析方法 [J]. 兰州理工大学学报,2022,48(3):94-102.

[33] DO H H,PRASAD P W C,MAAG A,et al. Deep Learning for Aspect-Based Sentiment Analysis:A Comparative Review [J].Expert Systems with Applications,2019,118:272-299.

[34] MINAEE S,KALCHBRENNER N,CAMBRIA E,et al. Deep Learning Based Text Classification:A Comprehensive Review [J/OL].arXiv:2004.03705 [cs.CL].[2022-08-11].https://arxiv.org/ abs/2004.03705.

[35] XU H,LIU B,SHU L,et al. BERT Post-Training for Review Reading Comprehension and Aspect-Based Sentiment Analysis [J/OL].arXiv:1904.02232 [cs.CL].[2022-08-12].https://arxiv.org/ abs/1904.02232.

[36] PHAN M H,OGUNBONA P O. Modelling Context and Syntactical Features for Aspect-Based Sentiment Analysis [C]// Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics.ACL,2020:3211-3220.

[37] RAMSHAW L A,MARCUS M P. Text Chunking Using Transformation-Based Learning [M]//ARMSTRONG S,CHURCH K, ISABELLE P. Natural Language Processing Using very Large Corpora. Dordrecht:Springer,1999:157-176.

[38] 王巍 . 解析网络舆情下官民互动的蝴蝶效应 [J]. 武警学院学报,2017,33(11):81-84.

[39] 于龙 . 涉警网络舆情危机的治理对策 [J]. 中国人民警察大学学报,2022,38(6):25-30.


作者简介:徐子昂(2001.03—),男,汉族,河南信阳人,本科在读,研究方向:自然语言处理。