A Comparative Empirical Study of Graph Neural Network Methods for Node-Level Financial Fraud Detection on Public Benchmark Graphs
Keywords:
graph neural networks, fraud detection, graph anomaly detection, benchmark evaluationAbstract
Graph neural networks have become a common tool for identifying fraudulent reviews, accounts, and transactions on relational platforms, yet published evaluations rarely share a common protocol, training ratio, or hyperparameter budget, making it difficult to judge which architectural family deserves credit for reported gains. This study consolidates results for nine representative graph neural network methods spanning three families, generic message-passing baselines, spatial camouflage-resistant detectors, and spectral heterophily-aware detectors, across four public benchmark graphs covering opinion fraud, transaction fraud, and anti-money-laundering tasks. Drawing on the original method papers and a large-scale open benchmark, the comparison finds that specialized camouflage-resistant detectors retain a moderate advantage over generic baselines on multi-relation review graphs under label-scarce training, while the same baselines close most of that gap given a larger, properly tuned training budget. On single-relation transaction graphs, spectral detectors and generic attention-based models perform comparably, with no single family holding a decisive advantage. Training ratio and hyperparameter tuning appear to explain a substantial share of the performance differences attributed to architectural novelty in prior work, and generalization across relation structures deserves more attention than raw accuracy tables typically give it. The comparison is assembled entirely from previously published and openly benchmarked figures, without new model training, so every reported value remains traceable to a public source.
References
Kipf, T. N., & Welling, M. (2017). Semi-supervised classification with graph convolutional networks. In Proceedings of the International Conference on Learning Representations.
Dou, Y., Liu, Z., Sun, L., Deng, Y., Peng, H., & Yu, P. S. (2020). Enhancing graph neural network-based fraud detectors against camouflaged fraudsters. In Proceedings of the 29th ACM International Conference on Information and Knowledge Management (pp. 315–324). Association for Computing Machinery.
Liu, Y., Ao, X., Qin, Z., Chi, J., Feng, J., Yang, H., & He, Q. (2021). Pick and choose: A GNN-based imbalanced learning approach for fraud detection. In Proceedings of the Web Conference 2021 (pp. 3168–3177). Association for Computing Machinery.
Huang, X., Yang, Y., Wang, Y., Wang, C., Zhang, Z., Xu, J., Chen, L., & Vazirgiannis, M. (2022). DGraph: A large-scale financial dataset for graph anomaly detection. In Advances in Neural Information Processing Systems 35, Datasets and Benchmarks Track.
Hamilton, W. L., Ying, Z., & Leskovec, J. (2017). Inductive representation learning on large graphs. In Advances in Neural Information Processing Systems 30.
Veličković, P., Cucurull, G., Casanova, A., Romero, A., Liò, P., & Bengio, Y. (2018). Graph attention networks. In Proceedings of the International Conference on Learning Representations.
Liu, Z., Dou, Y., Yu, P. S., Deng, Y., & Peng, H. (2020). Alleviating the inconsistency problem of applying graph neural network to fraud detection. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval (pp. 1569–1572). Association for Computing Machinery.
Shi, F., Cao, Y., Shang, Y., Zhou, Y., Zhou, C., & Wu, J. (2022). H2-FDetector: A GNN-based fraud detector with homophilic and heterophilic connections. In Proceedings of the ACM Web Conference 2022 (pp. 1486–1494). Association for Computing Machinery.
Huang, M., Liu, Y., Ao, X., Li, K., Chi, J., Feng, J., Yang, H., & He, Q. (2022). AUC-oriented graph neural network for fraud detection. In Proceedings of the ACM Web Conference 2022 (pp. 1311–1321). Association for Computing Machinery.
Wang, Y., Zhang, J., Huang, Z., Li, W., Feng, S., Ma, Z., Sun, Y., Yu, D., Dong, F., Jin, J., Wang, B., & Luo, J. (2023). Label information enhanced fraud detection against low homophily in graphs. In Proceedings of the ACM Web Conference 2023.
Duan, M., Zheng, T., Gao, Y., Wang, G., Feng, Z., & Wang, X. (2024). DGA-GNN: Dynamic grouping aggregation GNN for fraud detection. Proceedings of the AAAI Conference on Artificial Intelligence, 38.
Chai, Z., You, S., Yang, Y., Pu, S., Xu, J., Cai, H., & Jiang, W. (2022). Can abnormality be detected by graph neural networks? In Proceedings of the International Joint Conference on Artificial Intelligence.
Gao, Y., Wang, X., He, X., Liu, Z., Feng, H., & Zhang, Y. (2023). Addressing heterophily in graph anomaly detection: A perspective of graph spectrum. In Proceedings of the ACM Web Conference 2023.
Tang, J., Hua, F., Gao, Z., Zhao, P., & Li, J. (2023). GADBench: Revisiting and benchmarking supervised graph anomaly detection. In Advances in Neural Information Processing Systems 36, Datasets and Benchmarks Track.
Weber, M., Domeniconi, G., Chen, J., Weidele, D. K. I., Bellei, C., Robinson, T., & Leiserson, C. E. (2019). Anti-money laundering in Bitcoin: Experimenting with graph convolutional networks for financial forensics. KDD Workshop on Anomaly Detection in Finance.
Xiang, S., Zhu, M., Cheng, D., Li, E., Zhao, R., Ouyang, Y., Chen, L., & Zheng, Y. (2023). Semi-supervised credit card fraud detection via attribute-driven graph representation. Proceedings of the AAAI Conference on Artificial Intelligence, 37.
Zou, Y., & Cheng, D. (2024). Effective high-order graph representation learning for credit card fraud detection. In Proceedings of the International Joint Conference on Artificial Intelligence.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Innovative Technologies
This work is licensed under a Creative Commons Attribution 4.0 International License.