融合BERT多层次特征的短视频网络舆情情感分析研究

doi:10.3778/j.issn.1673-9418.2311023

摘要/Abstract

摘要： 自媒体时代与网络社交软件的广泛普及，导致短视频平台极易成为舆情事件起源和发酵的“孵化器”。分析短视频平台中的舆情评论信息，对于舆情事件的预警、处置和引导具有重要意义。鉴于此，结合BERT与TextCNN模型，提出一种融合BERT多层次特征的文本分类模型（BERT-MLFF-TextCNN），并对抖音短视频平台中的相关评论文本数据进行情感分析。首先，利用BERT预训练模型对输入文本进行编码。其次，提取各编码层中的语义特征向量进行融合。然后，融入自注意力机制突出其关键特征，从而实现特征的有效利用。最后，将所得特征序列输入TextCNN模型中进行分类。实验结果表明，与BERT-TextCNN、GloVe-TextCNN和Word2vec-TextCNN模型相比，BERT-MLFF-TextCNN模型表现更优，[F1]值达到了0.977。通过该模型能够有效识别短视频平台舆情的情感倾向，在此基础上利用TextRank算法进行主题挖掘，实现舆情评论情感极性的主题词可视化，为相关部门的舆情管控工作提供决策参考。

关键词: 网络舆情, 情感分析, 主题可视化, BERT

Abstract: The era of self-media and the widespread popularity of online social software have led to short video platforms becoming “incubators” easily for the origin and fermentation of public opinion events. Analyzing the public opinion comments on these platforms is crucial for the early warning, handling, and guidance of such incidents. In view of this, this paper proposes a text classification model combining BERT and TextCNN, named BERT-MLFF-TextCNN, which integrates multi-level features from BERT for sentiment analysis of relevant comment data on the Douyin short video platform. Firstly, the BERT pre-trained model is used to encode the input text. Secondly, semantic feature vectors from each encoding layer are extracted and fused. Subsequently, a self-attention mechanism is integrated to highlight key features, thereby effectively utilizing them. Finally, the resulting feature sequence is input into the TextCNN model for classification. The results demonstrate that the BERT-MLFF-TextCNN model outperforms BERT-TextCNN, GloVe-TextCNN, and Word2vec-TextCNN models, achieving an [F1] score of 0.977. This model effectively identifies the emotional tendencies in public opinions on short video platforms. Based on this, using the TextRank algorithm for topic mining allows for the visualization of thematic words related to the sentiment polarity of public opinion comments, providing a decision-making reference for relevant departments in the public opinion management work.

Key words: network public opinion, sentiment analysis, theme visualization, BERT

韩坤, 潘宏鹏, 刘忠轶. 融合BERT多层次特征的短视频网络舆情情感分析研究[J]. 计算机科学与探索, 2024, 18(4): 1010-1020.

HAN Kun, PAN Hongpeng, LIU Zhongyi. Research on Sentiment Analysis of Short Video Network Public Opinion by Integrating BERT Multi-level Features[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(4): 1010-1020.

参考文献

[1] 中国互联网络信息中心. 第52次《中国互联网络发展状况统计报告》[EB/OL]. [2023-08-28]. https://www.cnnic.cn/n4/2023/0828/c88-10829.html.
China Internet Network Information Center. The 52nd statistical report on China’s Internet development[EB/OL].[2023-08-28]. https://www.cnnic.cn/n4/2023/0828/c88-10829.html.
[2] 祁凯, 李昕. 基于S3EIR模型的政务短视频网民情绪感染路径研究[J]. 情报理论与实践, 2022, 45(10): 164-168.
QI K, LI X. Research on propagation path about netizens’ emotions of government affairs short video based on S3EIR model[J]. Information Studies: Theory & Application, 2022, 45(10): 164-168.
[3] 孙靖超. 基于优化深度双向自编码网络的舆情情感识别研究[J]. 情报杂志, 2020, 39(6): 159-163.
SUN J C. Sentiment analysis of network public opinion based on optimized bidirectional encoder representations from transformers[J]. Journal of Intelligence, 2020, 39(6): 159-163.
[4] 潘苏楠, 邓三鸿, 王蔚萍. 融合母评论文本信息的评论短文本情感分析研究——以短视频评论文本为例[J]. 情报探索, 2023(4): 1-7.
PAN S N, DENG S H, WANG W P. Research on the short text sentiment analysis with the text of parent-comment: a case study of short video comments[J]. Information Research, 2023(4): 1-7.
[5] SHI W, ZHANG J, HE S Y. Understanding public opinions on Chinese short video platform by multimodal sentiment analysis using deep learning based techniques[J/OL]. Kybernetes [2023-09-12]. https://www.emerald.com/insight/content/ doi/10.1108/K-04-2023-0723/full/html.
[6] 王婷, 杨文忠. 文本情感分析方法研究综述[J]. 计算机工程与应用, 2021, 57(12): 11-24.
WANG T, YANG W Z. Review of text sentiment analysis methods[J]. Computer Engineering and Applications, 2021, 57(12): 11-24.
[7] 周知, 李名子, 崔旭. 基于领域情感词典的用户生成内容有用性评价研究——以豆瓣读书为例[J]. 情报理论与实践, 2022, 45(1): 86-92.
ZHOU Z, LI M Z, CUI X. Research on helpfulness evaluation of user generate content based on domain sentiment lexicon: taking Douban reading as an example[J]. Information Studies: Theory & Application, 2022, 45(1): 86-92.
[8] 柳位平, 朱艳辉, 栗春亮, 等. 中文基础情感词词典构建方法研究[J]. 计算机应用, 2009, 29(10): 2875-2877.
LIU W P, ZHU Y H, LI C L, et al. Research on building Chinese basic semantic lexicon[J]. Journal of Computer Applications, 2009, 29(10): 2875-2877.
[9] WU L, MORSTATTER F, LIU H. SlangSD: building, expanding and using a sentiment dictionary of slang words for short-text sentiment classification[J]. Language Resources and Evaluation, 2018, 45(2): 196-211.
[10] ZHANG S, WEI Z, WANG Y, et al. Sentiment analysis of Chinese micro-blog text based on extended sentiment dictionary[J]. Future Generation Computer Systems, 2018, 81: 395-403.
[11] NEETHU M S, RAJASREE R. Sentiment analysis in Twitter using machine learning techniques[C]//Proceedings of the 4th International Conference on Computing, Jul 4-6, 2013. Piscataway: IEEE, 2013: 450-454.
[12] PANG B, LEE L, VAITHYANATHAN S. Thumbs up? Sentiment classification using machine learning techniques[C]//Proceedings of the 2002 Conference on Empirical Methods in Natural Language Processing, Philadelphia, Jul 6-7, 2002. New York: ACM, 2002: 79-86.
[13] JADAV B M, VAGHELA V B. Sentiment analysis using support vector machine based on feature selection and semantic analysis[J]. International Journal of Computer Applications, 2016, 146(13): 26-30.
[14] DEY L, CHAKRABORTY S, BISWAS A, et al. Sentiment analysis of review datasets using naive Bayes and K-NN classifier[J]. International Journal of Information Engineering and Electronic Business, 2016, 8(4): 54-62.
[15] 冯兴杰, 张志伟, 史金钏. 基于卷积神经网络和注意力模型的文本情感分析[J]. 计算机应用研究, 2018, 35(5): 1434-1436.
FENG X J, ZHANG Z W, SHI J C. Text sentiment analysis based on convolutional neural networks and attention model[J]. Application Research of Computers, 2018, 35(5): 1434-1436.
[16] TANG D, QIN B, LIU T. Document modeling with gated recurrent neural network for sentiment classification[C]//Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2015: 1422- 1432.
[17] DEVLIN J, CHANG M W, LEE K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[J]. arXiv:1810.04805, 2018.
[18] 刘思琴, 冯胥睿瑞. 基于BERT的文本情感分析[J]. 信息安全研究, 2020, 6(3): 220-227.
LIU S Q, FENG X R R. Text sentiment analysis based on BERT[J]. Journal of Information Security Research, 2020, 6(3): 220-227.
[19] 方英兰, 孙吉祥, 韩兵. 基于BERT的文本情感分析方法的研究[J]. 信息技术与信息化, 2020(2): 108-111.
FANG Y L, SUN J X, HAN B. Research on text emotion analysis method based on BERT[J]. Information Technology and Informatization, 2020(2): 108-111.
[20] 科猫网. 抖音用户有多少（抖音2021年数据分析）[EB/OL]. [2023-09-12]. http://news.kemaowang.org.cn/n/315774.html.
Kemao Net. How many Douyin users are there (data analysis of Douyin in 2021) [EB/OL]. [2023-09-12]. http://news.kemaowang.org.cn/n/315774.html.
[21] 李航. 统计学习方法[M]. 北京: 清华大学出版社, 2012.
LI H. Statistical learning methods[M]. Beijing: Tsinghua University Press, 2012.
[22] 罗婉丽, 张磊. 结合拓扑势与TextRank算法的关键词提取方法[J]. 计算机应用与软件, 2022, 39(1): 334-338.
LUO W L, ZHANG L. Keywords extraction method combining topological potential and TextRank algorithm[J]. Computer Applications and Software, 2022, 39(1): 334-338.
[23] 龙玥, 刘译阳. 新媒体环境下高校负面网络舆情传播特征和路径研究[J]. 情报科学, 2019, 37(12): 134-139.
LONG Y, LIU Y Y. Propagation characteristics and paths of negative network public opinions in colleges under the new media environment[J]. Information Science, 2019, 37(12): 134-139.
[24] 王国华, 曾润喜. 解码网络舆情[M]. 武汉: 华中科技大学出版社, 2006.
WANG G H, ZENG R X. Explore Internet public opinion[M]. Wuhan: Huazhong University of Science & Techonlogy Press, 2006.
[25] 李淑娜, 郭洪波. 网络舆情演变的情绪机制及干预策略研究[J]. 山东社会科学, 2022(8): 162-168.
LI S N, GUO H B. Research on the emotional mechanism and intervention strategies of the evolution of online public opinion[J]. Shandong Social Sciences, 2022(8): 162-168.