融合超图增强与双重对比学习的情感分析方法

doi:10.3778/j.issn.1002-8331.2408-0146

摘要/Abstract

摘要： 情感分析是自然语言处理领域中一项具有挑战性的任务。现有的情感分析方法在语句的建模过程中往往会忽略词汇在不同环境下的差异化表达。针对此类问题，提出一种基于超图增强结合双重对比学习的文本情感分析模型（hypergraph enhancement and dual contrastive learning，HDCL）。该模型通过标签感知的数据增强，强化情感的显式表达。使用超图结构学习文本语义之间的高阶潜在关联，并以双重对比学习优化语句表征，以实现丰富、准确的语句情感表示。该模型在真实数据集上的实验结果相较于其他基准模型有所提升，同时设计的消融实验与超参数分析进一步说明了提出的模型对于情感分析任务的有效性。

关键词: 文本情感分析, 超图, 双重对比学习

Abstract: Sentiment analysis is a fundamental challenge in natural language processing. However, existing methods often fail to capture the nuanced expressions of words across different contexts during sentence modeling. To address this limitation, this paper proposes a novel text sentiment analysis model HDCL that integrates hypergraph enhancement with dual contrastive learning. This model strengthens the explicit representation of sentiment through label-aware data augmentation. By employing a hypergraph structure, it captures high-order latent associations between text semantics, while dual contrastive learning is utilized to optimize sentence representations, resulting in richer and more accurate sentiment representation. Experimental results on real-world datasets demonstrate that the proposed model outperforms several benchmark models. Additionally, ablation studies and hyperparameter analysis further validate the effectiveness of the proposed model for sentiment analysis tasks.

Key words: text sentiment analysis, hypergraph, dual contrastive learning

孙帅祺, 魏桂英, 武森. 融合超图增强与双重对比学习的情感分析方法[J]. 计算机工程与应用, 2025, 61(22): 137-147.

SUN Shuaiqi, WEI Guiying, WU Sen. Sentiment Analysis Method Integrating Hypergraph Enhancement and Dual Contrastive Learning[J]. Computer Engineering and Applications, 2025, 61(22): 137-147.

参考文献

[1] AN J Y, WAN ZAINON W M N. Integrating color cues to improve multimodal sentiment analysis in social media[J]. Engineering Applications of Artificial Intelligence, 2023, 126: 106874.
[2] YANG X C, FENG S, WANG D L, et al. Image-text multimodal emotion classification via multi-view attentional network[J]. IEEE Transactions on Multimedia, 2021, 23: 4014-4026.
[3] ZHU L N, ZHU Z C, ZHANG C W, et al. Multimodal sentiment analysis based on fusion methods: a survey[J]. Information Fusion, 2023, 95: 306-325.
[4] 闫尚义, 王靖亚, 刘晓文, 等. 基于多头自注意力池化与多粒度特征交互融合的微博情感分析[J]. 数据分析与知识发现, 2023, 7(4): 32-45.
YAN S Y, WANG J Y, LIU X W, et al. Microblog sentiment analysis with multi-head self-attention pooling and multi-granularity feature interaction fusion[J]. Data Analysis and Knowledge Discovery, 2023, 7(4): 32-45.
[5] 吴旭旭, 陈鹏, 江欢. 基于多特征融合的微博细粒度情感分析[J]. 数据分析与知识发现, 2023, 7(12): 102-113.
WU X X, CHEN P, JIANG H. Micro-blog fine-grained sentiment analysis based on multi-feature fusion[J]. Data Analysis and Knowledge Discovery, 2023, 7(12): 102-113.
[6] FAN S, LIN C, LI H N, et al. Sentiment-aware word and sentence level pre-training for sentiment analysis[J]. arXiv:2210. 09803, 2022.
[7] WANKHADE M, RAO A C S, KULKARNI C. A survey on sentiment analysis methods, applications, and challenges[J]. Artificial Intelligence Review, 2022, 55(7): 5731-5780.
[8] WU F Z, HUANG Y F, SONG Y Q, et al. Towards building a high-quality microblog-specific Chinese sentiment lexicon[J]. Decision Support Systems, 2016, 87: 39-49.
[9] MOREO A, ROMERO M, CASTRO J L, et al. Lexicon-based comments-oriented news sentiment analyzer system[J]. Expert Systems with Applications, 2012, 39(10): 9166-9180.
[10] TRIPATHY A, AGRAWAL A, RATH S K. Classification of sentiment reviews using n-gram machine learning approach[J]. Expert Systems with Applications, 2016, 57: 117-126.
[11] WU F Z, HUANG Y F, SONG Y Q. Structured microblog sentiment classification via social context regularization[J]. Neurocomputing, 2016, 175: 599-609.
[12] AL AMRANI Y, LAZAAR M, EL KADIRI K E. Random forest and support vector machine based hybrid approach to sentiment analysis[J]. Procedia Computer Science, 2018, 127: 511-520.
[13] KIM Y. Convolutional neural networks for sentence classification[C]//Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2014: 1746-1751.
[14] LIU P F, QIU X P, HUANG X J. Recurrent neural network for text classification with multi-task learning[J]. arXiv:1605. 05101, 2016.
[15] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems 30, 2017: 5998-6008.
[16] YIN D, MENG T, CHANG K W. SentiBERT: a transferable transformer-based architecture for compositional sentiment semantics[J]. arXiv:2005.04114, 2020.
[17] TIAN H, GAO C, XIAO X Y, et al. SKEP: sentiment knowledge enhanced pre-training for sentiment analysis[J]. arXiv: 2005.05635, 2020.
[18] SUN Y D, ZHU D J, DU H W, et al. Motifs-based recommender system via hypergraph convolution and contrastive learning[J]. Neurocomputing, 2022, 512: 323-338.
[19] YAO L, MAO C S, LUO Y. Graph convolutional networks for text classification[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1): 7370-7377.
[20] LIN Y X, MENG Y X, SUN X F, et al. BertGCN: transductive text classification by combining GNN and BERT[C]//Fin-dings of the Association for Computational Linguistics: ACL-IJCNLP 2021. Stroudsburg: ACL, 2021: 1456-1462.
[21] WANG Y Z, WANG C X, ZHAN J Y, et al. Text FCG: fusing contextual information via graph learning for text classification[J]. Expert Systems with Applications, 2023, 219: 119658.
[22] PHAN H T, NGUYEN N T. A fuzzy graph convolutional network model for sentence-level sentiment analysis[J]. IEEE Transactions on Fuzzy Systems, 2024, 32(5): 2953-2965.
[23] LIU W X, ZHANG Z Z, WANG B. Dual-view hypergraph attention network for news recommendation[J]. Engineering Applications of Artificial Intelligence, 2024, 133: 108256.
[24] DING K Z, WANG J L, LI J D, et al. Be more with less: hypergraph attention networks for inductive text classification[J]. arXiv:2011.00387, 2020.
[25] YAN X D, SONG T W, JIAO Y F, et al. Spatio-temporal hypergraph learning for next POI recommendation[C]//Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2023: 403-412.
[26] NING Q, ZHAO Y M, GAO J, et al. AMHMDA: attention aware multi-view similarity networks and hypergraph learning for miRNA disease associations identification[J]. Briefings in Bioinformatics, 2023, 24(2): bbad094.
[27] LI J K, WANG J H, WU L, et al. AMHGCN: adaptive multi-level hypergraph convolution network for human motion prediction[J]. Neural Networks, 2024, 172: 106153.
[28] YUAN X, SUN M Y, CHEN Z K, et al. Semantic clustering-based deep hypergraph model for online reviews semantic classification in cyber-physical-social systems[J]. IEEE Access, 2018, 6: 17942-17951.
[29] 李全鑫, 庞俊, 朱峰冉. 结合Bert与超图卷积网络的文本分类模型[J]. 计算机工程与应用, 2023, 59(17): 107-115.
LI Q X, PANG J, ZHU F R. Text classification method based on integration of Bert and hypergraph convolutional network[J]. Computer Engineering and Applications, 2023, 59(17): 107-115.
[30] LE-KHAC P H, HEALY G, SMEATON A F. Contrastive representation learning: a framework and review[J]. IEEE Access, 2020, 8: 193907-193934.
[31] KLEIN T, NABI M. miCSE: mutual information contrastive learning for low-shot sentence embeddings[J]. arXiv:2211. 04928, 2022.
[32] LI J C, SHANG J B, MCAULEY J. UCTopic: unsupervised contrastive learning for phrase representations and topic mining[J]. arXiv:2202.13469, 2022.
[33] GIORGI J, NITSKI O, WANG B, et al. DeCLUTR: deep contrastive learning for unsupervised textual representations[J]. arXiv:2006.03659, 2020.
[34] PAN L, HANG C W, SIL A, et al. Improved text classification via contrastive adversarial training[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2022, 36(10): 11130-11138.
[35] WANG D, DING N, LI P J, et al. CLINE: contrastive learning with semantic negative examples for natural language understanding[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing. Stroudsburg: ACL, 2021: 2332-2342.
[36] KIM T, YOO K M, LEE S G. Self-guided contrastive learning for BERT sentence representations[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing. Stroudsburg: ACL, 2021: 2528-2540.
[37] JIAN Z Q, LI J J, WU Q Q, et al. Retrieval contrastive lear-ning for aspect-level sentiment classification[J]. Information Processing & Management, 2024, 61(1): 103539.
[38] DEVLIN J, CHANG M W, LEE K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[J]. arXiv:1810.04805, 2018.
[39] LIU Y H, OTT M, GOYAL N, et al. RoBERTa: a robustly optimized BERT pretraining approach[J]. arXiv:1907.11692, 2019.
[40] FENG Y F, YOU H X, ZHANG Z Z, et al. Hypergraph neural networks[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1): 3558-3565.
[41] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780.
[42] LI M, LONG Y, QIN L, et al. Emotion corpus construction based on selection from hashtags[C]//Proceedings of the 10th International Conference on Language Resources and Evaluation, 2016: 1845-1849.
[43] 胥桂仙, 刘兰寅, 王家诚, 等. 基于BERT和超图对偶注意力网络的文本情感分析[J]. 计算机应用研究, 2024, 41(3): 786-793.
XU G X, LIU L Y, WANG J C, et al. Text sentiment analysis based on BERT and hypergraph with dual attention network[J]. Application Research of Computers, 2024, 41(3): 786-793.
[44] 张佳威, 王中卿, 陈嘉沥. 基于文本生成的多粒度评论情感分析[J/OL]. 计算机科学 [2024-08-03]. https://link.cnki.net/urlid/50.1075.tp.20241122.1115.008.
ZHANG J W, WANG Z Q, CHEN J L. Multi-grained sentiment analysis of comments based on text generation[J/OL]. Computer Science [2024-08-03]. https://link.cnki.net/urlid/50.1075.tp.20241122.1115.008.
[45] CUI Y M, CHE W X, LIU T, et al. Pre-training with whole word masking for Chinese BERT[J]. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2021, 29: 3504-3514.
[46] LOSHCHILOV I, HUTTER F. Decoupled weight decay regularization[J]. arXiv:1711.05101, 2017.
[47] XIAO Z, LIANG P J. Chinese sentiment analysis using bi-directional LSTM with word embedding[C]//Proceedings of the 2nd International Conference on Cloud Computing and Security. Cham: Springer, 2016: 601-610.
[48] WANG J L, DING K Z, HONG L J, et al. Next-item recommendation with sequential hypergraphs[C]//Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2020: 1101-1110.