This article examines the possibilities of using machine learning methods in credit risk assessment for commercial banks. Based on the latest statistical data from Uzbekistan's banking sector, the study compares traditional credit scoring models with modern algorithms - Random Forest, XGBoost, and logistic regression. The findings indicate that ensemble methods provide superior predictive accuracy compared to traditional approaches. The article concludes with practical recommendations for improving credit risk management in the context of digital transformation of Uzbekistan's commercial banks.
Umida et al. (Tue,) studied this question.