基于交易序列与图结构的以太坊欺诈检测方法

doi:10.3969/j.issn.1671-1122.2026.09.013

信息网络安全 ›› 2026, Vol. 26 ›› Issue (9): 1481-1492.doi: 10.3969/j.issn.1671-1122.2026.09.013

基于交易序列与图结构的以太坊欺诈检测方法

黄佳欣¹, 刘睿远², 段荣成², 田野¹(), 孙建国¹

¹ 西安电子科技大学杭州研究院，杭州 311231
² 军工保密资格审查认证中心，北京 100089

收稿日期:2026-06-08 出版日期:2026-09-10 发布日期:2026-09-30
通讯作者: 田野 E-mail:tianye@xidian.edu.cn
作者简介:黄佳欣（2003—），女，浙江，硕士研究生，主要研究方向为网络安全、深度学习、以太坊|刘睿远（2001—），男，北京，助理工程师，本科，主要研究方向为数据安全|段荣成（1995—），男，山东，工程师，本科，主要研究方向为网络安全|田野（1992—），男，黑龙江，副教授，博士，CCF会员，主要研究方向为区块链安全、多模态信息处理、深度学习|孙建国（1981—），男，黑龙江，教授，博士，主要研究方向为网络安全、互联网安全、数据安全、深度学习
基金资助:
国家自然科学基金(62302365)

Ethereum fraud detection method based on transaction sequences and graph structures

Huang Jiaxin¹, Liu Ruiyuan², Duan Rongcheng², Tian Ye¹(), Sun Jianguo¹

¹ Hangzhou Institute of Technology, Xidian University, Hangzhou 311231, China
² Defence Industry Secrecy Examination and Certification Center, Beijing 100089, China

Received:2026-06-08 Online:2026-09-10 Published:2026-09-30
Contact: Tian Ye E-mail:tianye@xidian.edu.cn

摘要/Abstract

摘要：

以太坊正面临日益复杂的欺诈威胁，而现有网络钓鱼检测方法多依赖于原始交易网络的静态分析，忽视了账户细粒度的动态交易序列模式，未能充分挖掘结构与序列特征的协同效益。因此，文章提出一种 TransG4Eth框架，将交易级Transformer与图神经网络相结合，以同时捕获交易上下文信息与账户交互关系。首先，构建了基于 Transformer 的交易序列编码器，将账户的完整交易历史建模为细粒度行为特征；其次，设计上下文地址恢复预训练任务，引导模型学习交易序列中账户间的隐含交互模式；再次，构建全局账户交互图，以显式方式刻画结构拓扑；最后，提出隐-显式对偶学习框架，在嵌入级与输出级实现两阶段融合。实验结果表明，在真实数据集上TransG4Eth显著优于现有方法，F1分数达85.99%。

关键词: 区块链, 以太坊, 欺诈检测, Transformer, 图表示学习

Abstract:

Ethereum faces increasingly complex fraud threats, while existing phishing detection methods mostly relied on static analysis of raw transaction networks, overlooked fine-grained dynamic transaction sequence patterns of accounts, and failed to fully exploit the synergistic benefits of structural and sequential features. Therefore, this paper proposed a TransG4Eth framework, which combined a transaction-level Transformer with a graph neural network to capture both transaction context information and account interaction relationships. The proposed method first constructed a Transformer-based transaction sequence encoder, which modeled the complete transaction history of accounts as fine-grained behavioral features. Second, it designed a contextual address recovery pre-training task, which guided the model to learn implicit interaction patterns among accounts in transaction sequences. Then, it constructed a global account interaction graph to explicitly characterize structural topology. Finally, it proposed an implicit-explicit dual learning framework, which achieved two-stage fusion at both the embedding level and the output level. Experimental results show that TransG4Eth significantly outperforms existing methods on real-world datasets, obtains an F1 Score of 85.99%.

Key words: blockchain, Ethereum, fraud detection, Transformer, graph representation learning

中图分类号:

TP309

黄佳欣, 刘睿远, 段荣成, 田野, 孙建国. 基于交易序列与图结构的以太坊欺诈检测方法[J]. 信息网络安全, 2026, 26(9): 1481-1492.

Huang Jiaxin, Liu Ruiyuan, Duan Rongcheng, Tian Ye, Sun Jianguo. Ethereum fraud detection method based on transaction sequences and graph structures[J]. Netinfo Security, 2026, 26(9): 1481-1492.

图/表 8

图1

表1

表2

图2

表3

图3

表4

图4

参考文献 29

[1]	李鹏超, 张全涛, 胡源. 基于双注意力机制图神经网络的智能合约漏洞检测方法[J]. 信息网络安全, 2024, 24(11):1624-1631.
[2]	Wu Jiajing, Yuan Qi, Lin Dan, et al. Who are the phishers? Phishing scam detection on Ethereum via network embedding[J]. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2022, 52(2): 1156-1166. doi: 10.1109/TSMC.2020.3016821 URL
[3]	Hu Sihao, Zhang Zhen, Lu Shengliang, et al. Sequence-based target coin prediction for cryptocurrency pump-and-dump[J]. Proceedings of the ACM on Management of Data, 2023, 1(1): 1-19.
[4]	Chen Weili, Zheng Zibin, Cui Jiahui, et al. Detecting ponzi schemes on Ethereum: towards healthier blockchain technology[C]// The 2018 World Wide Web Conference. New York: ACM, 2018: 1409-1418.
[5]	Bian Shuqing, Deng Zhenpeng, Li Fei, et al. Icorating: a deep-learning system for scam ICO identification[EB/OL]. (2018-03-09)[2026-06-01]. https://arxiv.org/abs/1803.03670.
[6]	Victor F, Weintraud A M. Detecting and quantifying wash trading on decentralized cryptocurrency exchanges[C]// The Web Conference 2021. New York: ACM, 2021: 23-32.
[7]	Daian P, Goldfeder S, Kell T, et al. Flash boys 2.0: frontrunning, transaction reordering, and consensus instability in decentralized exchanges[EB/OL]. (2019-04-10)[2026-06-07]. https://arxiv.org/abs/1904.05234.
[8]	Sun Jianguo, Jia Yifan, Wang Yanbin, et al. Ethereum fraud detection via joint transaction language model and graph representation learning[J]. Information Fusion, 2025, 120: 1-12.
[9]	Zhang Sheng, Song Liangliang, Wang Yanbin. Dynamic feature fusion: combining global graph structures and local semantics for blockchain phishing detection[J]. IEEE Transactions on Network and Service Management, 2025, 22(5): 4706-4718. doi: 10.1109/TNSM.2025.3576130 URL
[10]	边玲玉, 张琳琳, 赵楷, 等. 基于LightGBM的以太坊恶意账户检测方法[J]. 信息网络安全, 2020, 20(4): 73-80.
[11]	Tian Ye, Song Liangliang, Qian Peng, et al. DiT-SGCR: directed temporal structural representation with global-cluster awareness for Ethereum malicious account detection[EB/OL]. (2025-06-25)[2026-06-07]. https://arxiv.org/abs/2506.20123.
[12]	Tan Runnan, Tan Qingfeng, Zhang Peng, et al. Graph neural network for Ethereum fraud detection[C]// 2021 IEEE International Conference on Big Knowledge (ICBK). New York: IEEE, 2021: 78-85.
[13]	Kanezashi H, Suzumura T, Liu Xin, et al. Ethereum fraud detection with heterogeneous graph neural networks[EB/OL]. (2022-03-23)[2026-06-07]. https://arxiv.org/abs/2203.12363.
[14]	Tan Runnan, Tan Qingfeng, Zhang Qin, et al. Ethereum fraud behavior detection based on graph neural networks[J]. Computing, 2023, 105: 2143-2170. doi: 10.1007/s00607-023-01177-7
[15]	Li Panpan, Xie Yunyi, Xu Xinyao, et al. Phishing fraud detection on Ethereum using graph neural network[EB/OL]. (2022-04-18)[2026-06-07]. https://doi.org/10.48550/arXiv.2204.08194.
[16]	Liu Jiayi, Yin Changchun, Wang Hao, et al. Graph embedding-based money laundering detection for Ethereum[J]. Electronics, 2023, 12(14): 1-14. doi: 10.3390/electronics12010001 URL
[17]	Hu Sihao, Zhang Zhen, Luo Binqiao, et al.BERT4ETH: a pre-trained Transformer for Ethereum fraud detection[C]// The ACM Web Conference 2023. New York: ACM, 2023: 2189-2197.
[18]	Hu Sihao, Huang Tiansheng, Chow K H, et al. ZipZap: efficient training of language models for large-scale fraud detection on blockchain[C]// The ACM Web Conference 2024. New York: ACM, 2024: 2807-2816.
[19]	Chen Zhen, Liu Shengzheng, Huang Jia, et al. Ethereum phishing scam detection based on data augmentation method and hybrid graph neural network model[J]. Sensors, 2024, 24(12): 1-17. doi: 10.3390/s24010001 URL
[20]	Wang Jinhuan, Chen Pengtao, Xu Xinyao, et al. TSGN: transaction subgraph networks assisting phishing detection in Ethereum[EB/OL]. (2022-08-27)[2026-06-07]. https://arxiv.org/abs/2208.12938.
[21]	Zhang Jiale, Sui Hao, Zhu Donghai, et al. GrabPhisher: phishing scams detection in Ethereum via temporally evolving GNNs[J]. IEEE Transactions on Services Computing, 2024, 17(6): 3727-3741. doi: 10.1109/TSC.2024.3411449 URL
[22]	Liu Lin, Tsai W T, Bhuiyan M Z A, et al. Blockchain-enabled fraud discovery through abnormal smart contract detection on Ethereum[J]. Future Generation Computer Systems, 2022, 128: 158-166. doi: 10.1016/j.future.2021.08.023 URL
[23]	Li Sijia, Gou Gaopeng, Liu Chang, et al. TTAGN: temporal transaction aggregation graph network for Ethereum phishing scams detection[C]// The ACM Web Conference 2022. New York: ACM, 2022: 661-669.
[24]	Lee S, Park T, Lee K. Learning to embed time series patches independently[EB/OL]. (2024-05-02)[2026-06-07]. https://doi.org/10.48550/arXiv.2312.16427.
[25]	Ekambaram V, Jati A, Nguyen N, et al. TSMixer: lightweight MLP-Mixer model for multivariate time series forecasting[C]// The 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: ACM, 2023: 459-469.
[26]	Ahmed N K, Rossi R A, Lee J B, et al.Role2Vec: role-based network embeddings[C]// The 2nd International Workshop on Deep Learning on Graphs: Methods and Applications. New York: ACM, 2019: 1-7.
[27]	Kipf T N, Welling M. Semi-supervised classification with graph convolutional networks[EB/OL]. (2016-09-09)[2026-06-07]. https://arxiv.org/abs/1609.02907.
[28]	Velickovic P, Cucurull G, Casanova A, et al. Graph attention networks[EB/OL]. (2017-10-30)[2026-06-07]. https://arxiv.org/abs/1710.10903.
[29]	Hamilton W L, Ying R, Leskovec J. Inductive representation learning on large graphs[EB/OL]. (2018-09-10)[2026-06-07]. https://doi.org/10.48550/arXiv.1706.02216.

实际账户类型	预测为正常账户	预测为钓鱼账户
正常账户	TN	FP
钓鱼账户	FN	TP

方法	Precision	Recall	F1-Score	B-Acc
RoleVec	57.48%	79.58%	66.73%	75.07%
Trans2Vec	26.34%	70.43%	38.42%	35.98%
GCN	55.15%	75.08%	63.59%	72.28%
GAT	47.29%	83.48%	60.38%	68.48%
SAGE	45.89%	58.26%	51.34%	61.96%
BERT4ETH	74.21%	61.25%	67.11%	75.30%
ZipZap	73.56%	61.12%	66.76%	75.06%
本文方法	84.42%	87.62%	85.99%	89.62%

模型	Precision	Recall	F1-Score	B-Acc
w/o 对偶学习	72.29%	90.62%	80.43%	86.30%
w/o 序列建模	79.81%	85.23%	82.43%	87.02%
TransG4Eth	84.42%	87.62%	85.99%	89.62%

λ	Precision	Recall	F1-Score	B-Acc
0	82.65%	93.21%	87.62%	91.53%
0.1	84.07%	91.62%	87.68%	91.31%
0.2	83.07%	93.01%	87.76%	91.59%
0.3	82.82%	91.42%	86.91%	90.79%
0.4	82.17%	92.02%	86.82%	90.83%
0.5	79.86%	91.02%	85.07%	89.56%
0.6	82.29%	89.02%	85.52%	89.54%
0.7	81.37%	88.02%	84.56%	88.78%
0.8	81.40%	88.22%	84.67%	88.88%
0.9	84.11%	88.22%	84.11%	88.52%
1.0	78.91%	86.63%	82.59%	87.31%

基于交易序列与图结构的以太坊欺诈检测方法

Ethereum fraud detection method based on transaction sequences and graph structures

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 29

相关文章 15

编辑推荐

Metrics

本文评价

[1]	张健, 蔡鸿, 王柏韬. 基于硬件性能计数器的恶意软件检测方法研究[J]. 信息网络安全, 2026, 26(9): 1383-1390.
[2]	朱晓强, 林彦孜, 张皓文, 姚英英. 基于多维特征与动态防御的区块链联邦学习鲁棒聚合方法[J]. 信息网络安全, 2026, 26(9): 1391-1407.
[3]	付章杰, 刘天怿, 宋甫元, 姜琴. 一个支持区块链分片的可验证跨链查询方案[J]. 信息网络安全, 2026, 26(8): 1209-1223.
[4]	韩妍妍, 杜雨芃, 余卓洲, 刘昂, 刘雯. 基于半量子网络的安全区块链[J]. 信息网络安全, 2026, 26(8): 1237-1249.
[5]	张宏涛, 王清涛, 张连成, 王吉昌. Argus6：一种IPv6 DDoS威胁检测模型[J]. 信息网络安全, 2026, 26(8): 1264-1276.
[6]	刘哲凡, 郭亚军, 郭奕旻, 赵素芬. 抗隐式攻击的基于ASCON的轻量级工业物联网认证方案[J]. 信息网络安全, 2026, 26(8): 1308-1326.
[7]	郑天明, 刘尚东, 李海天, 李华. 一种基于交互式卷积和Transformer的异常检测算法[J]. 信息网络安全, 2026, 26(7): 1077-1086.
[8]	李海龙, 张运豪, 沈燮阳, 邢宇航, 崔治安. 基于机器学习的恶意软件检测方法综述[J]. 信息网络安全, 2026, 26(4): 521-541.
[9]	李锦凯, 王靖雯, 董立波, 姚文翰, 刘成杰, 文伟平. 基于时序图注意力网络的区块链异常交易检测方法[J]. 信息网络安全, 2026, 26(4): 579-590.
[10]	陈宇琪, 钱汉伟, 夏玲玲, 王群. FEViT：一种基于频域增强ViT的深度伪造检测模型[J]. 信息网络安全, 2026, 26(3): 432-441.
[11]	罗文华, 许世龙, 吕涛, 张宁. 融合布谷鸟过滤与时序B+树的区块链数据溯源机制[J]. 信息网络安全, 2026, 26(3): 442-451.
[12]	梁广俊, 裘宇辰, 司宏韬, 王群, 马卓, 陈宇琪. 区块链虚拟货币溯源与分析[J]. 信息网络安全, 2026, 26(3): 482-490.
[13]	郭毅, 李旭青, 张子蛟, 张宏涛, 张连成, 张香丽. 基于区块链的数据安全共享研究[J]. 信息网络安全, 2026, 26(1): 1-23.
[14]	董佳瑜, 高宏民, 马兆丰, 赖冠辉. 多重签名机制在区块链中的研究及实现[J]. 信息网络安全, 2026, 26(1): 125-138.
[15]	赵文宇, 党晨曦, 杜振华, 张健. 基于硬件性能计数器的勒索软件检测技术研究与实现[J]. 信息网络安全, 2025, 25(9): 1397-1406.