Research on ML-based Multi-model Comparison & Feature Importance Analysis in Corporate Credit Scoring Models

In this study, a prediction model is constructed based on a comprehensive corporate information and business conditions dataset to predict the risk of corporate customers. The model will be applied to credit assessment, loan approval, risk management, and business intelligence analysis to provide scientific decision support for financial institutions and enterprises. The study explores the impact of features, deals with missing data and outliers, and compares different machine learning models to determine the best risk prediction model. The employment of these methodologies enables the effective identification and prevention of potential risks in enterprises, thereby enhancing their risk management capabilities.