Temporal Graph Learning for Default Prediction and Systemic Risk Mitigation in Financial Networks
Intelligent Computing · 2025-01 · 4
DOI: 10.34133/icomputing.0193 ↗Kurz gefasst
Fachartikel in der Zeitschrift Intelligent Computing, online erschienen am 13.11.2025. Verfasst von Moyang Liu und 4 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 63 Literaturangaben.
Abstract (Originalsprache)
In the realm of networked-loans, it is common for small and medium enterprises (SMEs) to obtain bank loans through guarantor companies. However, during economic downturns, the default of a few firms can trigger a cascade effect within the network, potentially resulting in a comprehensive system collapse. When a firm defaults, its guarantor is contractually required to settle the debt, which may strain the guarantor’s financial health and trigger further defaults across connected firms. This cascading effect can transform isolated financial distress into systemic risk. Therefore, the prediction of firms’ defaults from a comprehensive perspective is vital for ensuring the long-term stability of economic growth. To address this challenge, we propose a temporal graph learning framework that integrates macroeconomic trends and a tailored risk matrix, capturing both external shocks and internal contagion risks. Validated on a large commercial bank dataset with a total loan volume of USD 3 trillion, our system risk dynamic prediction (SRDP) method achieves 88.3% area under curve (AUC), considerably outperforming the state-of-the-art baselines. Risk mitigation analysis reveals that targeting the top 1% of predicted high-risk nodes reduces systemic risk exposure by an average of 25%, with heightened efficacy during financial crises and market volatility phases. The proposed method empowers financial regulators and risk managers with enhanced capabilities to monitor and grasp risks embedded within networked-loans against systemic financial crises.
Zitieren
APA
Liu, M., Li, T., Chen, J., Niu, Z., & Zhang, J. (2025). Temporal Graph Learning for Default Prediction and Systemic Risk Mitigation in Financial Networks. Intelligent Computing, 4. https://doi.org/10.34133/icomputing.0193
Vancouver
Liu M, Li T, Chen J, Niu Z, Zhang J. Temporal Graph Learning for Default Prediction and Systemic Risk Mitigation in Financial Networks. Intelligent Computing. 2025;4. doi:10.34133/icomputing.0193
BibTeX
@article{Liu2025,
author = {Liu, Moyang and Li, Taijun and Chen, Junqi and Niu, Zhibin and Zhang, Jiawan},
title = {Temporal Graph Learning for Default Prediction and Systemic Risk Mitigation in Financial Networks},
journal = {Intelligent Computing},
year = {2025},
volume = {4},
doi = {10.34133/icomputing.0193}
}
EUVIMED