Heterogeneous enhancement and multi-source knowledge fusion for aspect-based sentiment analysis

doi:10.11772/j.issn.1001-9081.2025080979

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (9): 2752-2760.DOI: 10.11772/j.issn.1001-9081.2025080979

• Artificial intelligence • Previous Articles

Heterogeneous enhancement and multi-source knowledge fusion for aspect-based sentiment analysis

Yifan WANG¹, Hu HAN²(), Dong LI², Yating FAN², Lin LI¹

^1.School of Mathematics and Physics，Lanzhou Jiaotong University，Lanzhou Gansu 730070，China
^2.School of Electronic and Information Engineering，Lanzhou Jiaotong University，Lanzhou Gansu 730070，China

Received:2025-08-25 Revised:2025-10-17 Accepted:2025-10-20 Online:2026-09-16 Published:2026-09-10
Contact: Hu HAN
About author:WANG Yifan， born in 1999， M. S. candidate. Her research interests include natural language processing， aspect-based sentiment analysis.
HAN Hu， born in 1977， Ph. D.， professor. His research interests include data mining， natural language processing.
LI Dong， born in 2000， M. S. candidate. His research interests include natural language processing， multimodal sentiment analysis.
FAN Yating， born in 1999， M. S. Her research interests include natural language processing.
LI Lin， born in 1998， M. S. Her research interests include emotional dictionary construction.
Supported by:
National Natural Science Foundation of China(62166024);Key Research and Development Program of Gansu Province - Industrial Category(25YFGA046)

基于异构增强与多源知识融合的方面级情感分析

王伊璠¹, 韩虎²(), 李栋², 范雅婷², 李琳¹

^1.兰州交通大学数理学院，兰州 730070
^2.兰州交通大学电子与信息工程学院，兰州 730070

通讯作者: 韩虎
作者简介:王伊璠（1999—），女，山西运城人，硕士研究生，主要研究方向：自然语言处理、方面级情感分析
韩虎（1977—），男，甘肃兰州人，教授，博士，主要研究方向：数据挖掘、自然语言处理
李栋（2000—），男，安徽亳州人，硕士研究生，主要研究方向：自然语言处理、多模态情感分析
范雅婷（1999—），女，甘肃兰州人，硕士，主要研究方向：自然语言处理
李琳（1998—），女，甘肃酒泉人，硕士，主要研究方向：情感词典构建。
基金资助:
国家自然科学基金资助项目(62166024);甘肃省重点研发计划-工业类项目(25YFGA046)

Abstract

Abstract:

Aspect-Based Sentiment Analysis （ABSA） is an important task in natural language processing. The existing methods generally adopt single-type nodes to model semantic associations between aspects and opinion terms， failing to distinguish the contributions of different node types to sentiment representation effectively. Moreover， current models usually construct syntactic trees using dependency parsers for semantic analysis， ignoring fine-grained semantic information of the word itself and implicit sentiment intensity modification relations in the text. To address the above issues， a Heterogeneous enhancement and Multi-Source knowledge fusion Graph Convolutional Network （HMS-GCN） model was proposed. First， a heterogeneous graph with differentiated constraints was constructed to capture semantic and sentiment information in sentences multi-dimensionally by using the relations of multi-type nodes and their corresponding edges in the heterogeneous graph， and importance of different types of nodes and their internal components were focused on by utilizing a dual-channel attention mechanism. Second， sememe knowledge was adopted to enhance the semantic representation of initial textual features， and hierarchical phrase structural information was integrated on the basis of the syntactic adjacency matrix， so as to improve the model’s syntactic dependency comprehension. Finally， the heterogeneous graph network information and enhanced knowledge features were fused for sentiment classification prediction. Experimental results on public datasets （Twitter， Lap14， Rest15， Rest16） show that compared to models such as ISSK-GCN （Integrating Syntax and Sentiment Knowledge Graph Convolutional Network）， the proposed model achieves the best results， validating the effectiveness of the differentiated node modeling and knowledge enhancement strategies.

Key words: Aspect-Based Sentiment Analysis (ABSA), heterogeneous graph, dual-channel attention mechanism, Graph Convolutional Network (GCN), sememe knowledge

摘要：

方面级情感分析（ABSA）是自然语言处理中的重要研究任务，现有方法普遍采用单一类型节点建模句子中方面词与意见词的语义关联，难以有效区分不同类型节点对情感表达的差异性贡献。此外，现有的模型通常利用句法依赖解析器构建句法树进行语义分析，忽视了单词本身所蕴含的细粒度语义信息及文本中隐含的情感强度修饰关系。针对上述问题，提出一种基于异构增强与多源知识融合的图卷积网络（HMS-GCN）模型。首先，构建基于差异化约束的异构图网络，利用异构图中的多类型节点及其相应边关系，多维度捕捉句子的语义情感信息，并通过双通道注意力机制关注不同类型节点及节点内部的重要程度；其次，利用义原知识对初始文本特征进行语义增强，并在句法邻接矩阵的基础上融合层级短语的结构信息，从而提升模型对句法依存关系的理解能力；最后，融合异构图网络信息和知识增强特征进行情感分类预测。在公开数据集Twitter、Lap14、Rest15和Rest16上的实验结果表明，相较于ISSK-GCN（Integrating Syntax and Sentiment Knowledge Graph Convolutional Network）等模型，所提模型的准确率取得了最优结果，验证了差异化节点建模与知识增强策略的有效性。

关键词: 方面级情感分析, 异构图, 双通道注意力机制, 图卷积网络, 义原知识

CLC Number:

TP391.1

Yifan WANG, Hu HAN, Dong LI, Yating FAN, Lin LI. Heterogeneous enhancement and multi-source knowledge fusion for aspect-based sentiment analysis[J]. Journal of Computer Applications, 2026, 46(9): 2752-2760.

王伊璠, 韩虎, 李栋, 范雅婷, 李琳. 基于异构增强与多源知识融合的方面级情感分析[J]. 《计算机应用》唯一官方网站, 2026, 46(9): 2752-2760.

Figures/Tables 11

Fig. 1 Aspect-based sentiment analysis example and its dependency tree

Fig. 2 Framework of HMS-GCN

Fig. 3 Heterogeneous graph and its adjacency matrix

Fig. 4 Hierarchical phrase structure tree

Tab. 1 Dataset statistics

数据集	$N p o s$		$N n e u$		$N n e g$
数据集	训练集	测试集	训练集	测试集	训练集	测试集
Twitter	1 561	173	3 127	346	1 560	173
Lap14	994	341	464	169	870	128
Rest15	912	326	36	34	256	182
Rest16	1 240	469	69	30	439	117

Tab. 1 Dataset statistics

数据集	$N p o s$		$N n e u$		$N n e g$
数据集	训练集	测试集	训练集	测试集	训练集	测试集
Twitter	1 561	173	3 127	346	1 560	173
Lap14	994	341	464	169	870	128
Rest15	912	326	36	34	256	182
Rest16	1 240	469	69	30	439	117

Tab. 2 Comparison experimental results of different models

类别	模型	Twitter		Lap14		Rest15		Rest16
类别	模型	Acc	M-F1	Acc	M-F1	Acc	M-F1	Acc	M-F1
基于语法	BERT	75.28	74.11	77.58	72.38	81.32	70.89	88.01	66.32
	ASGCN	72.15	70.40	75.55	71.05	79.89	61.89	88.99	67.48
	BERT4GCN	74.73	73.76	77.49	73.01	83.23	67.27	87.78	75.34
	AGCN	75.43	74.11	79.94	76.52	82.84	65.08	88.80	67.65
	DA-GCN	75.43	73.67	78.82	75.28	82.97	64.56	89.69	71.86
	AEFN	77.55	76.63	81.49	78.31	—	—	—	—
基于知识	SK-GCN	75.00	73.01	79.00	75.57	83.20	66.78	87.19	72.02
	EK-GCN	75.89	75.16	81.30	79.19	—	—	89.36	69.32
	IA-HiNET	75.88	75.36	79.45	76.57	—	—	—	—
	MSKFSA	75.00	74.03	77.90	74.14	81.55	67.47	90.10	71.99
	TCKGCN	77.39	76.30	81.39	79.80	84.25	69.53	91.47	76.71
基于多图	TextGT	77.70	76.45	81.33	78.71	84.19	69.73	91.07	77.29
	ASHGAT	—	—	79.98	76.58	83.57	71.15	90.75	77.57
	ISSK-GCN	77.14	74.96	80.23	75.49	83.14	67.08	90.12	73.24
	HMS-GCN	78.32	77.12	81.56	77.89	84.69	67.74	92.53	79.20

Tab. 3 Results of ablation experiments

模型	Twitter		Lap14		Rest15		Rest16
模型	Acc	M-F1	Acc	M-F1	Acc	M-F1	Acc	M-F1
HMS-GCN	78.32	77.12	81.56	77.89	84.69	67.74	92.53	79.20
w/o Dual-Att	77.93	76.36	80.97	76.97	84.05	67.21	92.03	78.69
w/o T-Att	78.07	76.45	81.23	77.09	84.17	67.52	92.07	78.83
w/o N-Att	78.04	76.61	81.16	77.33	84.18	67.49	92.12	78.80
w/o sememe	76.98	75.38	80.29	76.54	84.32	66.52	91.26	78.43

Fig. 5 Impact of different types of nodes on model performance

Fig. 6 Impact of different dependency tree enhancement strategies on model performance

Fig. 7 Impact of graph convolutional network layers on model performance

Fig. 8 Dual-channel attention visualization for aspect term “service”

References 37

[1]	Fei H， Ren Y， Zhang Y， et al. Nonautoregressive encoder-decoder neural framework for end-to-end aspect-based sentiment triplet extraction ［J］. IEEE Transactions on Neural Networks and Learning Systems， 2023， 34（9）： 5544-5556.
[2]	Pontiki M， Galanis D， Pavlopoulos J， et al. SemEval-2014 task 4： aspect based sentiment analysis ［C］// SemEval 2014. Stroudsburg： ACL， 2014： 27-35.
[3]	仲兆满，吕慧慧，张渝，等. 融合动态边权重图注意力网络的方面级情感分析［J］. 计算机科学与探索， 2025， 19（11）： 3083-3093.
	Zhong Zhaoman， Huihui Lyu， Zhang Yu， et al. Aspect-level sentiment analysis with integrated dynamic edge-weighted graph attention network ［J］. Journal of Frontiers of Computer Science and Technology， 2025， 19（11）： 3083-3093.
[4]	Tang D， Qin B， Feng X， et al. Effective LSTMs for target-dependent sentiment classification ［C］// COLING 2016： Technical Papers. ［S.l.］： The COLING 2016 Organizing Committee， 2016： 3298-3307.
[5]	闫金凤，邵新慧. 基于图卷积网络的特定方面情感分析［J］. 中文信息学报， 2022， 36（10）： 135-144.
	Yan Jinfeng， Shao Xinhui. Aspect-level sentiment analysis based on graph convolutional network ［J］. Journal of Chinese Information Processing， 2022， 36（10）： 135-144.
[6]	Ouyang J， Xuan C， Wang B， et al. Aspect-based sentiment classification with aspect-specific hypergraph attention networks［J］. Expert Systems with Applications， 2024， 248： No.123412.
[7]	Gu T， Zhao H， Li M. Effective inter-aspect words modeling for aspect-based sentiment analysis ［J］. Applied Intelligence， 2023， 53（4）： 4366-4379.
[8]	Hao J， Pei L， He Y， et al. TCKGCN： graph convolutional network for aspect-based sentiment analysis with three-channel knowledge fusion ［J］. Neurocomputing， 2024， 600： No.128163.
[9]	Wang Y， Huang M， Zhu X， et al. Attention-based LSTM for aspect-level sentiment classification ［C］// EMNLP 2016. Stroudsburg： ACL， 2016： 606-615.
[10]	Ma D， Li S， Zhang X， et al. Interactive attention networks for aspect-level sentiment classification ［C］// IJCAI 2017. California： IJCAI， 2017： 4068-4074.
[11]	Zhang C， Li Q， Song D. Aspect-based sentiment classification with aspect-specific graph convolutional networks ［C］// EMNLP-IJCNLP 2019. Stroudsburg： ACL， 2019： 4568-4578.
[12]	Wang K， Shen W， Yang Y， et al. Relational graph attention network for aspect-based sentiment analysis ［C］// ACL 2020. Stroudsburg： ACL， 2020： 3229-3238.
[13]	Zhang M， Qian T. Convolution over hierarchical syntactic and lexical graphs for aspect level sentiment analysis ［C］// EMNLP 2020. Stroudsburg： ACL， 2020： 3540-3549.
[14]	Li R， Chen H， Feng F， et al. Dual graph convolutional networks for aspect-based sentiment analysis ［C］// ACL-IJCNLP 2021 （Volume 1： Long Papers）. Stroudsburg： ACL， 2021： 6319-6329.
[15]	Zhang Z， Zhou Z， Wang Y. SSEGCN： syntactic and semantic enhanced graph convolutional network for aspect-based sentiment analysis ［C］// NAACL-HLT 2022. Stroudsburg： ACL， 2022： 4916-4925.
[16]	Wang Y， Yang N， Miao D， et al. Dual-channel and multi-granularity gated graph attention network for aspect-based sentiment analysis ［J］. Applied Intelligence， 2023， 53（11）： 13145-13157.
[17]	Gu T， He Z， Zhao H， et al. Aspect-based sentiment analysis with multi-granularity information mining and sentiment hint ［J］. Expert Systems with Applications， 2024， 252（Pt A）： No.124104.
[18]	Yu B， Cao C， Yang Y. Dynamic position weighting aspect-focused graph convolutional network for aspect-based sentiment analysis［J］. The Journal of Supercomputing， 2025， 81： No.341.
[19]	Cambria E， Li Y， Xing F Z， et al. SenticNet 6： ensemble application of symbolic and subsymbolic ai for sentiment analysis［C］// CIKM 2020. New York： ACM， 2020： 105-114.
[20]	Xing F Z， Pallucchini F， Cambria E. Cognitive-inspired domain adaptation of sentiment lexicons ［J］. Information Processing and Management， 2019， 56（3）： 554-564.
[21]	Liang B， Su H， Gui L， et al. Aspect-based sentiment analysis via affective knowledge enhanced graph convolutional networks［J］. Knowledge-Based Systems， 2022， 235： No.107643.
[22]	Zhou J， Huang J， Hu Q V， et al. SK-GCN： modeling syntax and knowledge via graph convolutional network for aspect-level sentiment classification ［J］. Knowledge-Based Systems， 2020， 205： No.106292.
[23]	Zhong Q， Ding L， Liu J， et al. Knowledge graph augmented network towards multiview representation learning for aspect-based sentiment analysis ［J］. IEEE Transactions on Knowledge and Data Engineering， 2023， 35（10）： 10098-10111.
[24]	Cui X， Tao W， Cui X. Affective-knowledge-enhanced graph convolutional networks for aspect-based sentiment analysis with multi-head attention ［J］. Applied Sciences， 2023， 13（7）： No.4458.
[25]	韩虎，郝俊，张千锟，等. 多源知识融合的方面级情感分析模型［J］. 北京航空航天大学学报， 2024， 50（9）： 2688-2695.
	Han Hu， Hao Jun， Zhang Qiankun， et al. Multi-source knowledge fusion model for aspect-based sentiment analysis ［J］. Journal of Beijing University of Aeronautics and Astronautics， 2024， 50（9）： 2688-2695.
[26]	Zhu C， Ding Q. Aspect-based sentiment analysis via dual residual networks with sentiment knowledge ［J］. The Journal of Supercomputing， 2025， 81： No.131.
[27]	Devlin J， Chang M W， Lee K， et al. BERT： pre-training of deep bidirectional Transformers for language understanding ［C］// ACL NAACL-HLT 2019， Volume 1 （Long and Short Papers）. Stroudsburg： ACL， 2019： 4171-4186.
[28]	Qi P， Zhang Y， Zhang Y， et al. Stanza： a Python natural language processing toolkit for many human languages ［C］// ACL 2020： System Demonstrations. Stroudsburg： ACL， 2020： 101-108.
[29]	Dong Z， Dong Q. HowNet - a hybrid language and knowledge resource ［C］// 2003 International Conference on Natural Language Processing and Knowledge Engineering. Piscataway： IEEE， 2003： 820-824.
[30]	田继帅，艾芳菊. 基于增强句法信息与多特征图卷积融合的方面级情感分析［J］. 计算机科学与探索， 2025， 19（3）： 738-748.
	Tian Jishuai， Ai Fangju. Aspect-level sentiment analysis based on enhanced syntactic information and multi-feature graph convolutional fusion［J］. Journal of Frontiers of Computer Science and Technology， 2025， 19（3）： 738-748.
[31]	Xiao Z， Wu J， Chen Q， et al. BERT4GCN： using BERT intermediate layers to augment GCN for aspect-based sentiment classification ［C］// EMNLP 2021. Stroudsburg： ACL， 2021： 9193-9200.
[32]	Zhao M， Yang J， Zhang J， et al. Aggregated graph convolutional networks for aspect-based sentiment classification ［J］. Information Sciences， 2022， 600： 73-93.
[33]	Wang X， Liu P， Zhu Z， et al. Aspect-based sentiment analysis with graph convolutional networks over dependency awareness［C］// ICPR 2022. Piscataway： IEEE， 2022： 2238-2245.
[34]	郑诚，陈雪灵. 方面语义增强的融合网络用于方面级情感分析［J］. 小型微型计算机系统， 2025， 46（9）： 2105-2112.
	Zheng Cheng， Chen Xueling. Aspect semantic enhanced fusion network for aspect-based sentiment analysis ［J］. Journal of Chinese Computer Systems， 2025， 46（9）： 2105-2112.
[35]	Gu T， Zhao H， He Z， et al. Integrating external knowledge into aspect-based sentiment analysis using graph neural network［J］. Knowledge-Based Systems， 2023， 259： No.110025.
[36]	Yin S， Zhong G. TextGT： a double-view graph transformer on text for aspect-based sentiment analysis ［J］. Proceedings of the AAAI Conference on Artificial Intelligence， 2024， 38（17）： 19404-19412.
[37]	李自亮，朱广丽，张玉雷，等. 集成句法与情感知识的方面级情感分析模型［J］. 计算机应用， 2025， 45（6）： 1724-1731.
	Li Ziliang， Zhu Guangli， Zhang Yulei， et al. Aspect-based sentiment analysis model integrating syntax and sentiment knowledge ［J］. Journal of Computer Applications， 2025， 45（6）： 1724-1731.

Heterogeneous enhancement and multi-source knowledge fusion for aspect-based sentiment analysis

基于异构增强与多源知识融合的方面级情感分析

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 11

References 37

Related Articles 15

Recommended Articles

Metrics

[1]	Musheng CHEN, Wenqing FU, Xiaohong QIU, Junhua WU, Qiang WEN. Aspect sentiment triplet extraction based on graph convolutional network and cross-domain data augmentation [J]. Journal of Computer Applications, 2026, 46(8): 2432-2439.
[2]	Huihui LYU, Zhaoman ZHONG, Yu ZHANG, Jidong FAN. Syntactically enhanced aspect-based sentiment analysis via mutual information maximization and contrastive learning [J]. Journal of Computer Applications, 2026, 46(8): 2440-2446.
[3]	Pengyu CHEN, Baojun TIAN, Lichang ZHAO, Jiandong FANG. Learning path recommendation model integrating multi-behavior modeling and reinforcement learning [J]. Journal of Computer Applications, 2026, 46(8): 2485-2493.
[4]	Haoran YUAN, Huan LIU, Pengfei JIAO, Zhidong ZHAO, Xianfei ZHANG, Zunliang LIU. Masked autoencoder enhanced dynamic heterogeneous graph representation learning model [J]. Journal of Computer Applications, 2026, 46(6): 1728-1737.
[5]	Huijie GUO, Tianfeng DOU, Zhenlin ZHANG, Kaiyuan QI, Dong WU, Zhijian QU, Zhao LI, Chongguang REN. Time-interdependency-aware dynamic Bayesian network for traffic prediction [J]. Journal of Computer Applications, 2026, 46(5): 1507-1517.
[6]	Dirui ZHANG, Jiayu LIN, Zuhong LIANG. Supervised contrastive generative sentiment analysis method with uncertainty-aware unlikelihood learning [J]. Journal of Computer Applications, 2026, 46(5): 1416-1423.
[7]	Baoyuan ZHENG, Chaobo HE. Graph convolutional network enhanced by graph diffusion and dual-view feature learning [J]. Journal of Computer Applications, 2026, 46(5): 1370-1377.
[8]	Wenhao LI, Yinzhang GUO. Urban traffic flow prediction based on dual-layer multi-scale dynamic graph convolutional network model [J]. Journal of Computer Applications, 2026, 46(4): 1323-1333.
[9]	Chao SHI, Yuxin ZHOU, Qian FU, Wanyu TANG, Ling HE, Yuanyuan LI. Action recognition algorithm for ADHD patients using skeleton and 3D heatmap [J]. Journal of Computer Applications, 2025, 45(9): 3036-3044.
[10]	Biao ZHAO, Yuhua QIN, Rongkun TIAN, Yuehang HU, Fangrui CHEN. Dependency type and distance enhanced aspect based sentiment analysis model [J]. Journal of Computer Applications, 2025, 45(8): 2507-2514.
[11]	Jiaxin YAN, Yanping CHEN, Weizhe YANG, Ruizhang HUANG, Yongbin QIN. Heterogeneous graph attention network for relation extraction based on feature combination [J]. Journal of Computer Applications, 2025, 45(8): 2470-2476.
[12]	Ziliang LI, Guangli ZHU, Yulei ZHANG, Jiajia LIU, Yixuan JIAO, Shunxiang ZHANG. Aspect-based sentiment analysis model integrating syntax and sentiment knowledge [J]. Journal of Computer Applications, 2025, 45(6): 1724-1731.
[13]	Wenjing YAN, Ruidong WANG, Min ZUO, Qingchuan ZHANG. Recipe recommendation model based on hierarchical learning of flavor embedding heterogeneous graph [J]. Journal of Computer Applications, 2025, 45(6): 1869-1878.
[14]	Jian SHUAI, Zhongqing WANG, Jiali CHEN. Aspect-based sentiment analysis method based on code generation [J]. Journal of Computer Applications, 2025, 45(6): 1827-1832.
[15]	Quan WANG, Qixiang LU, Pei SHI. Multi-graph diffusion attention network for traffic flow prediction [J]. Journal of Computer Applications, 2025, 45(5): 1472-1479.