MCATSA: Multi-Strategy Collaborative Adaptive Tree-Seed Algorithm and Its Application in Dynamic Credit Risk Assessment

Kaiyu Huang

European Journal of AI, Computing & Informatics, 2026, vol. 2, issue 2, 61-75

Abstract: Credit risk assessment lies at the absolute core of maintaining global financial system stability and preventing systemic economic crises. However, traditional static evaluation models increasingly suffer from significant limitations when dealing with the complexities of multi-source heterogeneous data, intricate spatiotemporal risk contagion, and the urgent need for dynamic decision-making in modern financial markets. To comprehensively address these critical issues, this paper proposes a novel integrated risk control framework, denoted as IOA-DGNN-RL. This advanced architecture is fundamentally based on a newly improved Multi-Strategy Collaborative Adaptive Tree-Seed Algorithm (MCATSA), a Dynamic Graph Neural Network (DGNN), and Reinforcement Learning (RL) techniques. Firstly, the proposed MCATSA substantially improves the standard Tree-Seed Algorithm through the integration of five core mechanisms, which include heterogeneous chaotic initialization, multi-layer resource allocation, and nonlinear energy regulation. These enhancements achieve highly accurate dimensionality reduction of high-dimensional non-financial features, thereby optimizing computational efficiency. Secondly, a sophisticated DGNN model incorporating key macroeconomic indicators is constructed to precisely capture spatiotemporal risk propagation within the complex topological network among interconnected enterprises. Finally, a robust reinforcement learning decision module is designed to seamlessly transform dynamic risk predictions into actionable, optimal credit adjustment strategies. Comprehensive experiments demonstrate that the MCATSA performs excellently in standard benchmark optimization tasks. Furthermore, the complete integrated system achieves an impressive Area Under the Curve (AUC) of 0.901 on complex real-world credit datasets, significantly outperforming existing baseline methods and providing a highly reliable tool for modern financial risk management.

Keywords: tree-seed algorithm; graph neural networks; reinforcement learning; feature selection; credit risk assessment (search for similar items in EconPapers)
Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://pinnaclepubs.com/index.php/EJACI/article/view/746/717 (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:dba:ejacia:v:2:y:2026:i:2:p:61-75

Access Statistics for this article

More articles in European Journal of AI, Computing & Informatics from Pinnacle Academic Press
Bibliographic data for series maintained by Joseph Clark ().