Application of profit-based credit scoring models using R

Selçuk Bayracı (
R&D Centre, C/S Information Technologies, Istanbul, Turkey


In this study, we applied a profit-based scoring system with using 10 different statistical and machine learning algorithms on a consumer credit data of a Turkish commercial bank. RStudio environment and R packages have been used in data cleaning, feature selection and model implementation processes. The results of the study reveal that artificial neural networks model seems to be superior to other techniques in terms of profit maximization.

Keywords: Data analytics, Credit scoring, Banking, Risk management
JEL Classifcation: C01, C40, C87

[Full Text]

Romanian Statistical Review 4/2017