Contagion Path Prediction in Financial Enterprise Networks Using a Starling Murmuration Optimized Dual-Encoder Self-Attention GNN (SM-ISAGNN)
Abstract
Financial risk contagion refers to the cascading spread of financial distress among interconnected entities within a networked system, posing serious threats to economic stability. Predicting the pathways through which such contagion propagates is essential for early intervention and systemic risk mitigation. This research proposes a novel contagion path prediction algorithm based on a deep learning (DL) framework, designed to capture both the direct and indirect transmission of financial risks across enterprise networks. The Starling Murmuration Optimizer-driven improved Self-Attention Graph Neural Network (SM-ISAGNN) is applied to predict financial risk contagion paths in enterprise networks. The model was enhanced by integrating a dual-encoder architecture. The first encoder captures intra-entity risk using statistically significant financial indicators, legal records, and operational data. The second encoder models contagion dynamics using enterprise relation information derived from an enterprise knowledge graph. To collect the data, financial statements, including income, debt ratio, and liquidity indicators, along with credit scores and enterprise relationship information, were gathered from relevant financial databases and corporate records. The data was preprocessed by handling missing values and normalizing features. The SMO metaheuristic is employed to optimize attention weights, enhancing convergence and avoiding local minima. Experiments on a financial enterprise dataset demonstrate that SM-ISAGNN outperforms the baseline ISAGNN, achieving a higher path hit ratio (84.3% vs. 58.7%), a multi-hop detection rate (77.5% vs. 41.6%), and a lower false path prediction (7.9% vs. 21.3%). In 5-fold cross-validation, the model achieves an accuracy of 0.9667, a precision of 0.9658, a recall of 0.9667, and an F1-score of 0.9657. These results confirm SM-ISAGNN as a robust framework for early warning, risk visualization, and contagion path forecasting in financial enterprise networks.
Full Text:
PDFReferences
Chen, J. and Sun, B., 2024. Enhancing Financial Risk Prediction Using TG-LSTM Model: An Innovative Approach with Applications to Public Health Emergencies. Journal of the Knowledge Economy, pp.1-21. https://doi.org/10.1007/s13132-024-02081-x
Adeloye, F.C. and Olawoyin, O.M., 2025. Advanced financial derivatives in managing systemic risk and liquidity shocks in interconnected global markets. https://doi.org/10.30574/ijsra.2025.15.2.1500
Aliano, M., Cananà, L., Ciano, T., Ragni, S. and Ferrara, M., 2024. On the dynamics of a SIR model for a financial risk contagion. Quality & Quantity, pp.1-25. https://doi.org/10.1007/s11135-024-02009-2
Fialkowski, J., Diem, C., Borsos, A. and Thurner, S., 2025. A data-driven econo-financial stress-testing framework to estimate the effect of supply chain networks on financial systemic risk. arXiv preprint arXiv:2502.17044. https://doi.org/10.48550/arXiv.2502.17044
Dong, Y. and Dong, Z., 2023. An innovative approach to analyze financial contagion using causality-based complex network and value at risk. Electronics, 12(8), p.1846. https://doi.org/10.3390/electronics12081846
Vyas, A., 2025. Revolutionizing Risk: The Role of Artificial Intelligence in Financial Risk Management, Forecasting, and Global Implementation. Forecasting, and Global Implementation (April 21, 2025). https://dx.doi.org/10.2139/ssrn.5224657
Bello, O.A., 2023. Machine learning algorithms for credit risk assessment: an economic and financial analysis. International Journal of Management, 10(1), pp.109-133. https://doi.org/10.37745/ijmt.2013/vol10n1109133
Elhoseny, M., Metawa, N., Sztano, G. and El-Hasnony, I.M., 2025. Deep learning-based model for financial distress prediction. Annals of operations research, 345(2), pp.885-907. https://doi.org/10.1007/s10479-022-04766-5
Jia, K. and Pan, Y., 2025. Financial Systemic Risk Prediction Using Deep Neural Networks and Long Short-Term Memory. Journal of Circuits, Systems and Computers. https://doi.org/10.1142/S0218126625502949
Yang, T., Li, A., Xu, J., Su, G. and Wang, J., 2024. Deep learning model-driven financial risk prediction and analysis. https://doi.org/10.20944/preprints202406.2069.v1
Jin, Q., Sun, L., Chen, Y. and Hu, Z.L., 2024. Financial risk contagion based on dynamic multi-layer network between banks and firms. Physica A: Statistical Mechanics and its Applications, 638, p.129624. https://doi.org/10.1016/j.physa.2024.129624
Fan, R., Xie, X., Wang, Y. and Lin, J., 2025. Effect of financial contagion between real and financial sectors on asset bubbles: A two‐layer network game approach. Managerial and Decision Economics, 46(1), pp.393-408. https://doi.org/10.1002/mde.4381
Cheng, D., Niu, Z., Li, J. and Jiang, C., 2022. Regulating systemic crises: Stemming the contagion risk in networked-loans through deep graph learning. IEEE Transactions on Knowledge and Data Engineering, 35(6), pp.6278-6289. https://doi.org/10.1109/TKDE.2022.3162339
Chung, V., Espinoza, J. and Mansilla, A., 2024. Analysis of Financial Contagion and Prediction of Dynamic Correlations During the COVID-19 Pandemic: A Combined DCC-GARCH and Deep Learning Approach. Journal of Risk and Financial Management, 17(12), p.567. https://doi.org/10.3390/jrfm17120567
Liao, X. and Li, W., 2025. Research on the tail risk contagion in the international commodity market on the China's financial market: based on a network perspective. Kybernetes, 54(2), pp.807-831. https://doi.org/10.1108/K-06-2023-1001
Xie, X., Zhang, F., Liu, L., Yang, Y. and Hu, X., 2023. Assessment of associated credit risk in the supply chain based on trade credit risk contagion. PloS One, 18(2), p.e0281616. https://doi.org/10.1371/journal.pone.0281616
Dong, P., 2025. Research on Comprehensive Financial Crisis Warning Model of Listed Companies Based on Financial Report Big Data and Deep Learning Technology. The Frontiers of Society, Science and Technology, 7(2). https://doi.org/10.25236/FSST.2025.070205
Yao, Q., Mao, C. and Guo, Y., 2024. A Fuzzy Neural Network-Based Intelligent Warning Method for Financial Risk of Enterprises. Journal of Circuits, Systems and Computers, 33(14), p.2450251. https://doi.org/10.1142/S0218126624502517
Mu, P., Chen, T., Pan, K. and Liu, M., 2021. A Network Evolution Model of Credit Risk Contagion between Banks and Enterprises Based on Agent‐Based Model. Journal of Mathematics, 2021(1), p.6593218. https://doi.org/10.1155/2021/6593218
Ma, J., Liu, Y., Zhao, L. and Liang, W., 2024. Research on the mechanism and application of spatial credit risk contagion based on complex network model. Managerial and Decision Economics, 45(2), pp.1180-1193. https://doi.org/10.1002/mde.4025
Wang, L., Jiang, X., Chen, T. and Zhu, R., 2024. The Contagion of Debt Default Risk in Energy Enterprises Considering Carbon Price Fluctuations. Mathematics, 12(17), p.2776. https://doi.org/10.3390/math12172776
Li, M. and Fu, Y., 2022. Prediction of supply chain financial credit risk based on PCA-GA-SVM model. Sustainability, 14(24), p.16376. https://doi.org/10.3390/su142416376
Chen, X., 2025. Research on financial market volatility prediction and risk response strategy based on LSTM network. J. COMBIN. MATH. COMBIN. COMPUT, 127, pp.5197-5213. https://doi.org/10.61091/jcmcc127a-293
Kadkhoda, S.T. and Amiri, B., 2024. A hybrid network analysis and machine learning model for enhanced financial distress prediction. IEEE Access, 12, pp.52759-52777. https://doi.org/10.1109/ACCESS.2024.3387462
Ionescu, Ș., Delcea, C. and Nica, I., 2025. Improving Real-Time Economic Decisions Through Edge Computing: Implications for Financial Contagion Risk Management. Computers, 14(5), p.196. https://doi.org/10.3390/computers14050196
DOI: https://doi.org/10.31449/inf.v49i22.10485
This work is licensed under a Creative Commons Attribution 3.0 License.