多通道句法门控图神经网络用于句子级情感分析

doi:10.3778/j.issn.1002-8331.2311-0408

摘要/Abstract

摘要： 情感分析技术是自然语言处理领域的一项重要任务。然而，现阶段文档级图神经网络的图构建复杂且需要占用大量的内存资源。在线评论文本一般由短句组成，文档级图神经网络进行情感分析的效率较低。此外，现有工作中句子级图神经网络未能充分结合文本的单词特征、依存特征和词性特征。针对以上问题，提出一种多通道句法门控图神经网络的句子级情感分析方法（MSGNN）。该模型以句子的依存句法关系图为骨架，词性特征、单词特征和依存特征作为节点特征信息；利用三通道的门控图神经网络分别学习三种特征；采用图卷积神经网络聚合节点的特征信息。在SST-1、SST-2、MR三种基准数据集上的实验结果表明该模型相比基线模型的性能有所提升。

关键词: 情感分析, 句子级图神经网络, 依存特征, 门控图神经网络

Abstract: Sentiment analysis technology is an important task in the field of natural language processing. However, the graph construction of document-level graph neural network at this stage is complicated and requires a lot of memory resources. The online review text generally consists of short sentences, and document-level graph neural network for sentiment analysis is less efficient. Additionally, existing work on sentence-level graph neural networks has failed to adequately integrate the word features, dependency features, and part-of-speech features of the text. To address the above problems, this paper proposes a multi-channel syntactic gated graph neural network for sentence-level sentiment analysis (MSGNN). The model takes the dependency syntactic relation graph of sentences as the skeleton, and part-of-speech features, word features and dependency features as node feature information. Then a three-channel gated graph neural network is used to learn the three features separately. Finally a graph convolutional neural network is used to aggregate the feature information of the nodes. The experiment results on the three benchmark datasets, namely SST-1, SST-2, and MR, show that the performance of this model is improved compared with that of the baseline model.

Key words: sentiment analysis, sentence-level graph neural network, dependency features, gated graph neural network

张吴波, 邹旺, 熊黎, 戴顺鄂, 吴文欢. 多通道句法门控图神经网络用于句子级情感分析[J]. 计算机工程与应用, 2025, 61(8): 135-144.

ZHANG Wubo, ZOU Wang, XIONG Li, DAI Shun’e, WU Wenhuan. Multi-Channel Syntactic Gated Graph Neural Network for Sentence-Level Sentiment Analysis[J]. Computer Engineering and Applications, 2025, 61(8): 135-144.

参考文献

[1] HU M, LIU B. Mining and summarizing customer reviews[C]//Proceedings of the 10th ACM International Conference on Knowledge Discovery and Data Mining, 2004: 168-177.
[2] MEE A, HOMAPOUR E, CHICLANA F, et al. Sentiment analysis using TF-IDF weighting of UK MPs’ tweets on Brexit[J]. Knowledge-Based Systems, 2021, 228: 107238.
[3] SILVA D N F F, HRUSCHKA E R, HRUSCHKA J R E R. Tweet sentiment analysis with classifier ensembles[J]. Decision Support Systems, 2014, 66: 170-179.
[4] MULLEN T, COLLIER N. Sentiment analysis using support vector machines with diverse information sources[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2004: 412-418.
[5] LI Z, LI R, JIN G. Sentiment analysis of danmaku videos based on na?ve bayes and sentiment dictionary[J]. IEEE Access, 2020, 8: 75073-75084.
[6] NAGARAJAN S M, GANDHI U D. Classifying streaming of Twitter data based on sentiment analysis using hybridization[J]. Neural Computing and Applications, 2019, 31: 1425-1433.
[7] REHMAN A U, MALIK A K, RAZA B, et al. A hybrid CNN-LSTM model for improving accuracy of movie reviews sentiment analysis[J]. Multimedia Tools and Applications, 2019, 78: 26597-26613.
[8] WANG X, JIANG W, LOU Z. Combination of convolutional and recurrent neural network for sentiment analysis of short texts[C]//Proceedings of the 26th International Conference on Computational Linguistics: Technical Papers, 2016: 2428-2437.
[9] WANG B, LIANG B, DU J, et al. SEMGraph: incorporating sentiment knowledge and eye movement into graph model for sentiment analysis[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2022: 7521-7531.
[10] LI Y, ZEMEL R, BROCKSCHMIDT M, et al. Gated graph sequence neural networks[C]//Proceedings of the Conference on International Conference on Learning Representations, 2016.
[11] ZOU W, ZHANG W, WU W, et al. A multi-task shared cascade learning for aspect sentiment triplet extraction using BERT-MRC[J]. Cognitive Computation, 2024, 16(4): 1554-1571.
[12] MA Y, PENG H, KHAN T, et al. Sentic LSTM: a hybrid network for targeted aspect-based sentiment analysis[J]. Cognitive Computation, 2018, 10: 639-650.
[13] WANG J, ZHANG Y, YU L C, et al. Contextual sentiment embeddings via bi-directional GRU language model[J]. Knowledge-Based Systems, 2022, 235: 107663.
[14] ZOU W, ZHANG W, TIAN Z, et al. A hybrid model for text classification using part-of-speech features[J]. Journal of Intelligent & Fuzzy Systems, 2023, 45(1): 1-15.
[15] WANG J, YU L C, LAI K R, et al. Dimensional sentiment analysis using a regional CNN-LSTM model[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, 2016: 225-230.
[16] KIM Y. Convolutional neural networks for sentence classification[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2014: 1746-1751.
[17] LI W, ZHU L, SHI Y, et al. User reviews: sentiment analysis using lexicon integrated two-channel CNN-LSTM family models[J]. Applied Soft Computing, 2020, 94: 106435.
[18] CHEN T, XU R, HE Y, et al. Improving sentiment analysis via sentence type classification using BiLSTM-CRF and CNN[J]. Expert Systems with Applications, 2017, 72: 221-230.
[19] CHEN X, SUN C, WANG J, et al. Aspect sentiment classification with document-level sentiment preference modeling[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 2020: 3667-3677.
[20] ZHOU T, LAW K, CREIGHTON D. A weakly-supervised graph-based joint sentiment topic model for multi-topic sentiment analysis[J]. Information Sciences, 2022, 609: 1030-1051.
[21] PHAN H T, NGUYEN N T, HWANG D. Convolutional attention neural network over graph structures for improving the performance of aspect-level sentiment analysis[J]. Information Sciences, 2022, 589: 416-439.
[22] JIN Z, TAO M, ZHAO X, et al. Social media sentiment analysis based on dependency graph and co-occurrence graph[J]. Cognitive Computation, 2022, 14(3): 1039-1054.
[23] LIAO W, ZENG B, LIU J, et al. Multi-level graph neural network for text sentiment analysis[J]. Computers & Electrical Engineering, 2021, 92: 107096.
[24] DOZAT T, MANNING C D. Deep biaffine attention for neural dependency parsing[J]. arXiv:1611.01734, 2016.
[25] SOCHER R, PERELYGIN A, WU J, et al. Recursive deep models for semantic compositionality over a sentiment treebank[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2013: 1631-1642.
[26] PANG B, LEE L. Seeing stars: exploiting class relationships for sentiment categorization with respect to rating scales[C]//Proceedings of the Annual Meeting on Association for Computational Linguistics, 2005: 115-124.
[27] JOULIN A, GRAVE E, BOJANOWSKI P, et al. Bag of tricks for efficient text classification[C]//Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics, 2016: 427-431.
[28] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems, 2017: 6000-6010.
[29] DEVLIN J, CHANG M W, LEE K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[C]//Proceedings of the Conference on North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2019: 4171-4186.
[30] HUANG L, MA D, LI S, et al. Text level graph neural network for text classification[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, 2019: 3444-3450.
[31] ZHANG Y, YU X, CUI Z, et al. Every document owns its structure: inductive text classification via graph neural networks[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 2020: 334-339.
[32] LI Y, LI N. Sentiment analysis of Weibo comments based on graph neural network[J]. IEEE Access, 2022, 10: 23497-23510.
[33] ALBADANI B, SHI R, DONG J, et al. Transformer-based graph convolutional network for sentiment analysis[J]. Applied Sciences, 2022, 12(3): 1316.
[34] AN W, TIAN F, CHEN P, et al. Aspect-based sentiment analysis with heterogeneous graph neural network[J]. IEEE Transactions on Computational Social Systems, 2022, 10(1): 403-412.
[35] ZHAO M, YANG J, SHANG F. Dependency-enhanced graph convolutional networks for aspect-based sentiment analysis[J]. Neural Computing and Applications, 2023, 35(19): 14195-14211.