Credit risk analysis using support vector machines algorithm

Abstract

Credit risk or also known as bad credit risk is defined as the risk that occurs due to the inability of the customer to repay the loan and the interest within a certain period of time. Bad credit can be prevented by selecting proper customers during the loan application process so as not to give losses to the Bank. This research aims to analyze credit risk using machine learning methods. One of the machine learning algorithms that can be used for data classification is Support Vector Machine (SVM). The concept of SVM is to find optimal hyper plane with maximum margin to linearly separate the data set into two classes. The dataset were taken randomly from 2015 to 2018 at Bank XX, as many as 610 data. Dataset was divided into two parts along with percentage of the training data 80% and the testing data 20%. The variables used were gender, plafond, rate, term of time, job, income, face amount, warranty, and loan history as independent variable as well as credit status as dependent variable. The testing of SVM algorithm used linear, polynomial, Radial Basis Function (RBF), and sigmoid kernel obtained confusion matrices with accuracy respectively 0.9262, 0.9508, 0.8934, and 0.8361. Meanwhile, the AUC values were 0.9129, 0.9419, 0.9051, and 0.8285. The SVM model with polynomial kernel is the best model of the four models because it has the highest accuracy and AUC value. Thus, this model can be used to classify prospective customers into good credit or bad credit class with sufficiently high accuracy so as to help banks reduce the risk of bad credit.