Research on Supply Chain Payment Risk Identification and Prediction Methods Based on Machine Learning

Supply chain payment risk management has become increasingly critical in modern global commerce, where financial disruptions can cascade through entire networks of suppliers and manufacturers. Traditional risk assessment methodologies often fail to capture the dynamic and complex nature of contemporary supply chain environments, leading to substantial financial losses and operational disruptions. This research proposes a comprehensive machine learning-based framework for identifying and predicting supply chain payment risks through advanced feature engineering and ensemble learning techniques. The study develops a multi-dimensional risk assessment model that integrates supplier financial health indicators, transaction pattern analysis, and macroeconomic variables to enhance prediction accuracy. Experimental validation using real-world procurement and payment data demonstrates significant improvements in risk detection capabilities, achieving 94.2% accuracy in payment default prediction and reducing false positive rates by 37% compared to conventional methods. The proposed framework provides actionable insights for supply chain financial managers and contributes to the advancement of AI-driven risk management solutions in enterprise environments.