Artificial Intelligence Techniques for Risk prediction in financial management of listed companies based on optimized Deformable graph convolutional networks under digital economy: Trends and Challenges

Abstract

The rapid evolution of the digital economy has significantly increased the complexity of financial risk prediction for publicly listed companies, driven by interconnected markets, high-dimensional data, and real-time information flows. Traditional statistical models are often inadequate for capturing nonlinear dependencies and dynamic relationships inherent in modern financial systems, necessitating advanced artificial intelligence approaches. This paper presents a comprehensive review of graph-based deep learning methods, with a focus on Deformable Graph Convolutional Networks (DGCNs) for financial risk prediction. DGCNs enhance conventional graph neural networks by introducing adaptive receptive fields, enabling dynamic modeling of evolving relationships among financial entities such as firms, sectors, and markets. The review examines key optimization strategies including attention mechanisms, multi-scale feature fusion, residual connections, and contrastive learning to improve model performance and generalization. It also highlights the integration of heterogeneous data sources, including financial statements, market indicators, and sentiment-based features, to enrich predictive capabilities. Empirical studies across global financial datasets demonstrate that DGCN-based models outperform traditional and baseline methods in predicting credit risk, financial distress, and market volatility. Despite these advancements, challenges remain in scalability, interpretability, and regulatory compliance. This review provides insights into current methodologies and outlines future research directions for developing robust, scalable, and intelligent financial risk prediction systems in the digital economy.