Credit Risk Evaluation Model of Small-Micro Enterprises for Rural Commercial Bank Based on XGBoost and Random Forest

As an important pillar of the real economy, small-micro enterprises are facing many difficulties in operation under the dual impact of the global economic downturn and the COVID-19 epidemic. And they are in urgent need of external financing to tide over the difficulties. The purpose of rural commercial bank is to serve small and medium-sized enterprises and agriculture, rural areas and farmers. Therefore, the financing of rural commercial bank is crucial for the development of small-micro enterprises. However, due to unstable development mode, imperfect management mechanism and other reasons, small-micro enterprises usually have high credit risk, which brings great challenges to the loan risk control of rural commercial bank. Effective evaluation of credit risk of small-micro enterprises is the basis of rural commercial bank's loan decision-making and an important issue concerned by theory and practice. This paper studies the credit risk evaluation model of rural commercial bank for small-micro enterprises. Firstly, a set of credit risk evaluation indicator system combining financial and behavioral indicators is constructed, which expands the disadvantages of traditional research that only considers financial indicators is too one-sided. Secondly, XGBoost model is used to screen indicators and build a credit risk evaluation model of small-micro enterprises based on improved random forest. Finally, the effectiveness of the model is verified by comparing with other traditional models, and the experiment proves that the introduction of behavioral indicators can significantly improve the effectiveness of the model.