Credit Risk Assessment Model Based on XGBoost

Against the backdrop of uncertainty increasing global economic, Credit risk has become one of the core risks faced by financial institutions, enterprises and individuals. This study focuses on credit risk assessment in the field of financial risk and introduces machine learning models to address issues such as poor model interpretability and imbalanced data in existing models. By processing the German credit dataset in various aspects, including handling missing values, label encoding, standardization, data balancing, and analyzing the basic characteristics of the data to discover the advantages of the tree model on this model. The results indicate that the XGBoost model performs well in evaluation metrics such as accuracy, precision, recall, F1-score, and AUC. Compared to other traditional machine learning models, it has more significant advantages and can effectively identify credit risk. In addition, it was verified through ablation experiments that the model performance was the best after using the SMOTEENN data sampling method.