Early Warning of Companies' Credit Risk Based on Machine Learning

Abstract

With the advent of the big data era, information barriers are gradually being broken down and credit has become a key factor of company operations. The lack of company credit has greatly and negatively impacted the social economy, which has triggered considerable research on company credit. In this article, a credit risk warning model based on the XGBoost-SHAP algorithm is proposed that can accurately assess the credit risk of a company. The degree of influence of the characteristics of a company's credit risk and the warning threshold of important characteristics are obtained based on the model output. Finally, a comparison with several other machine learning algorithms showed that the XGBoost-SHAP model achieved the highest early warning accuracy and the most comprehensive explanatory output results. The experimental results show that the method can effectively provide a warning of the credit risk of a company based on the historical performance of the company's historical characteristics data. This method provides positive guidance for companies and financial institutions.