Algoritmos para Machine Learning utilizados en la Gestión de Riesgo Crediticio en Perú Machine Learning Algorithms used in Credit Risk Management in Peru

Abstract

 The purpose of this study is to identify the most used and effective Machine Learning algorithms in credit risk management in the Peruvian financial system. Four research studies were analyzed that implemented different algorithms, such as Decision Tree, Support Vector Machine (SVM), Random Forest, Neural Networks, Naive Bayes, K Nearest Neighbor and Logistic Regression, in financial and microfinance institutions. The methodology included a literature review and webgraphy. The results revealed that the most used algorithms were Decision Tree and SVM. In addition, Neural Networks and other algorithms presented acceptable performance, but did not outperform the leading models in credit risk prediction. The conclusions point out that, being Decision Tree and SVM the most effective options for predicting defaults and managing credit risk, the choice of the optimal model may depend on the specific characteristics of the data and above all the objectives of each financial institution.