Credit risk unveiled: Decision trees triumph in comparative machine learning study

Abstract

As times go on, credit risk has become a widespread issue across society, especially after the 2008 global financial crisis. However, the traditional financial technique could not determine the possibility of people defaulting, causing credit problems. With the rapid development of the Artificial Intelligence field, this could not be the problem. In this paper, several methods, including the Support Vector Machine model (SVM), K-Nearest Neighbors model (KNN) and Decision Tree model (DTs) are implemented using machine learning to try to predict the credit risk accurately and compare the accuracy of the three different methods. As a result, the Decision Trees show the highest result in these three methods.