Credit Risk Model Based on Central Bank Credit Registry Data

Data science and machine-learning techniques help banks to optimize enterprise operations, enhance risk analyses gain competitive advantage. There is a vast amount of research in credit risk, but our knowledge, none them uses registry as data source model the probability default for individual clients. The goal this paper evaluate different models create accurate assessment using from real dataset Central Bank Republic North Macedonia. We strongly believe that developed will be an additional valuable information commercial banks, by leveraging historical all population country banks. Thus, research, we compare five classify data, i.e., logistic regression, decision tree, random forest, support vector machines (SVM) neural network. metrics, propose based on central bank with detailed methodology can predict history country. Our results show best accuracy achieved tree performing imbalanced without scaling, followed forest linear regression.