Optimization of Scoring for SME Credit Risk Management via Machine Learning

Tari, Mohamed Larbi; Bouguerra, Amel; Selatni, Iatidel

Revue des sciences commerciales et de gestion
Volume 20, Numéro 1, Pages 31-60
2024-12-31

Optimization Of Scoring For Sme Credit Risk Management Via Machine Learning

Authors : Tari Mohamed Larbi . Bouguerra Amel . Selatni Iatidel .

Abstract

In the contemporary banking sector, managing credit default risk remains a pivotal challenge for financial institutions, especially in light of rapidly advancing technologies and increasingly complex economic conditions. Traditional evaluation methods, though insightful, often fall short in addressing the multifaceted nature of modern risk factors. This study investigates the application of Machine Learning techniques to enhance predictive models for credit risk, with a particular focus on small and medium-sized enterprises (SMEs). Drawing on a dataset of 647 credit applications and 17 explanatory variables, the research adopts an empirical approach. Four supervised learning algorithms—LR, DT, ANN, and RF—were implemented in R Studio software and assessed using confusion matrices and ROC curves. The findings highlight the Random Forest algorithm's superior predictive accuracy, underscoring the transformative role of technological innovations in optimizing risk management practices.