Comprehensive Credit Rating Method for Civil Aviation Supply Chain Enterprises Based on Dual-Model Fusion

Abstract

Traditional enterprise credit rating models have limitations such as insufficient industry adaptability and singular risk indicators, which make them inapplicable to the credit rating of civil aviation supply chain enterprises in business financing scenarios. This paper proposes a comprehensive credit rating model that integrates financial, litigation, operational, and industry indicators. By improving the weighted random forest algorithm (introducing SMOTE data balancing and decision tree weight optimization) and combining it with Logistic regression in a dual-model fusion, a dynamic scoring card system is constructed. The model strengthens short-term liquidity indicators (with a weight of 35%) in line with the characteristics of civil aviation supply chain enterprises. Empirical results show that the model achieves an average accuracy of 94.7% in five-fold cross-validation and increases the recall rate of bad samples to 96.2%, significantly improving the precision of risk assessment. This model has been applied to the financing scenario of civil aviation supply chain and can effectively alleviate the information asymmetry between banks and enterprises, providing reliable credit decision support for financial institutions.