A KAN-enhanced graphSAGE model for ethereum account classification on heterophilic graphs
Peer-to-Peer Networking and Applications · 22.06.2026 · 19 (4)
DOI: 10.1007/s12083-026-02223-9 ↗Kurz gefasst
Fachartikel in der Zeitschrift Peer-to-Peer Networking and Applications, online erschienen am 22.06.2026. Verfasst von Hengliang Guo und 6 weiteren Autor:innen. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: The research was supported by Science and Technology Innovation 2030 - ’Next-Generation Artificial Intelligence’ Major Project, Air pollution early warning based on domestic heterogeneous intelligent computing platform, Research on system architecture and application algorithms for the integration of supercomputing and quantum computing.
Abstract (Originalsprache)
Ethereum account classification is essential for identifying individuals engaged in illicit transactions and analyzing behavioral patterns across various account types. This process serves as a critical mechanism for monitoring and regulating unlawful activities within transactional markets. However, the Ethereum network exhibits the characteristics of a complex heterophilic graph which poses significant challenges to the effectiveness and performance of conventional graph neural networks (GNNs). To address this challenge, the present study proposes FSGCN(Fourier-Sage GCN), a novel architecture for heterophilic graph neural networks (GNNs) that integrates Kolmogorov–Arnold Networks (KANs) with GraphSAGE. FSGCN is specifically designed to adapt efficiently to the structural complexity of heterophilic graphs. By leveraging KANs to extract high-order neighborhood information and employing GraphSAGE to capture low-order neighborhood patterns, FSGCN effectively aggregates both homophilic and heterophilic features, thereby improving classification performance. Furthermore, to improve training efficiency and generalization, we propose the MLPInit weight initialization scheme and the DropEdge graph augmentation technique. Experiments on a large-scale Ethereum transaction dataset show that FSGCN achieves an F1-score of 91.8% and a classification accuracy of 91.6%, significantly outperforming traditional homophilic and heterophilic GNN baselines. Additionally, FSGCN demonstrates high training efficiency, completing each epoch in just 2.302 s per epoch and improving overall training speed by 130.4% compared to conventional GraphSAGE.
Förderung
- The research was supported by Science and Technology Innovation 2030 - ’Next-Generation Artificial Intelligence’ Major Project
- Air pollution early warning based on domestic heterogeneous intelligent computing platform
- Research on system architecture and application algorithms for the integration of supercomputing and quantum computing
- Postgraduate Education Reform and Quality Improvement Project of Henan Province
Zitieren
APA
Guo, H., Sui, Y., Li, J., Gao, F., Han, Y., Guo, Y., & Wu, G. (2026). A KAN-enhanced graphSAGE model for ethereum account classification on heterophilic graphs. Peer-to-Peer Networking and Applications, 19(4). https://doi.org/10.1007/s12083-026-02223-9
Vancouver
Guo H, Sui Y, Li J, Gao F, Han Y, Guo Y, et al. A KAN-enhanced graphSAGE model for ethereum account classification on heterophilic graphs. Peer-to-Peer Networking and Applications. 2026;19(4). doi:10.1007/s12083-026-02223-9
BibTeX
@article{Guo2026,
author = {Guo, Hengliang and Sui, Yizhe and Li, Jiaru and Gao, Fuchang and Han, Yubo and Guo, Yang and Wu, Gang},
title = {A KAN-enhanced graphSAGE model for ethereum account classification on heterophilic graphs},
journal = {Peer-to-Peer Networking and Applications},
year = {2026},
volume = {19},
number = {4},
doi = {10.1007/s12083-026-02223-9}
}
EUVIMED