Hybrid Graph Neural Network-Based Aspect-Level Sentiment Classification

被引:0
作者
Zhao, Hongyan [1 ]
Cui, Cheng [1 ]
Wu, Changxing [2 ]
机构
[1] Yang En Univ, Coll Engn, Quanzhou 362014, Peoples R China
[2] East China Jiaotong Univ, Sch Informat & Software Engn, Nanchang 330013, Peoples R China
基金
中国国家自然科学基金;
关键词
aspect-level sentiment classification; hybrid graph neural network; dependency syntax information; sentiment relationships;
D O I
10.3390/electronics13163263
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Aspect-level sentiment classification has received more and more attention from both academia and industry due to its ability to provide more fine-grained sentiment information. Recent studies have demonstrated that models incorporating dependency syntax information can more effectively capture the aspect-specific context, leading to improved performance. However, existing studies have two shortcomings: (1) they only utilize dependency relations between words, neglecting the types of these dependencies, and (2) they often predict the sentiment polarity of each aspect independently, disregarding the sentiment relationships between multiple aspects in a sentence. To address the above issues, we propose an aspect-level sentiment classification model based on a hybrid graph neural network. The core of our model involves constructing several hybrid graph neural network layers, designed to transfer information among words, between words and aspects, and among aspects. In the process of information transmission, our model takes into account not only dependency relations and their types between words but also sentiment relationships between aspects. Our experimental results based on three commonly used datasets demonstrate that the proposed model achieves a performance that is comparable to or better than recent benchmark methods.
引用
收藏
页数:17
相关论文
共 46 条
[1]  
Bo Pang, 2008, Foundations and Trends in Information Retrieval, V2, P1, DOI 10.1561/1500000001
[2]  
Bruna J., 2014, 2 INT C LEARN REPR I
[3]  
Chen CH, 2022, PROCEEDINGS OF THE 60TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2022), VOL 1: (LONG PAPERS), P2051
[4]  
Chen P., 2017, P 2017 C EMP METH NA, P452, DOI [10.18653/v1/D17-1047, DOI 10.18653/V1/D17-1047]
[5]  
Devlin J, 2019, 2019 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: HUMAN LANGUAGE TECHNOLOGIES (NAACL HLT 2019), VOL. 1, P4171
[6]  
Dozat Timothy, 2017, 5 INT C LEARN REPR I
[7]  
[杜成玉 Du Chengyu], 2020, [中文信息学报, Journal of Chinese Information Processing], V34, P70
[8]   An introduction to ROC analysis [J].
Fawcett, Tom .
PATTERN RECOGNITION LETTERS, 2006, 27 (08) :861-874
[9]   Integrating external knowledge into aspect-based sentiment analysis using graph neural network [J].
Gu, Tiquan ;
Zhao, Hui ;
He, Zhenzhen ;
Li, Min ;
Ying, Di .
KNOWLEDGE-BASED SYSTEMS, 2023, 259
[10]   Fusing sentiment knowledge and inter-aspect dependency based on gated mechanism for aspect-level sentiment classification [J].
Han, Yu ;
Zhou, Xiaotang ;
Wang, Guishen ;
Feng, Yuncong ;
Zhao, Hui ;
Wang, Junhua .
NEUROCOMPUTING, 2023, 551