Prediction of Loan Status Using Logistics Regression Model and Naïve Bayes Classifier

Abstract

Conducting an evaluation process of prospective debtors is important for creditors to reduce the risk of default. For this reason, the research aims to construct a model that can determine whether a prospective applicant's credit application is recommended to be accepted or rejected by using the method of logistic regression and naïve Bayes classifier. We used a dataset of gender, married, dependent, education, self-employed, applicant income, co-applicant income, loan amount, loan amount term, credit history, and property area as predictor variables and loan status as a response variable. The results show that the performance measures, including accuracy, precision, recall, and F1 score of the logistics regression method, are 85.9%, 83.82%, 100%, and 91.2%, while the naïve Bayes classifier is 84.62%, 83.58%, 98.2%, and 90.32%. Since the performance measures of logistic regression are bigger than naïve Bayes classifier, it suggests that logistic regression is better than naïve Bayes classifier.