Bank Loan Default Prediction Using Ensemble Machine Learning Algorithm

ABSTRACT

Banks play an integral role in the financial system of any country which directly affects its economic status and growth. The major roles of banks include accepting deposits from its customers, using those deposits to lend money to the borrowers in return for some interest, granting credits, discounting on bills etc. But the main source of profit for the banks is the interest it receives from lending money to the borrowers. And in a scenario of global pandemic like Covid-19, the number of people requiring financial aid from the banks has increased drastically. But a major problem faced by these banks is the failure of timely loan repayment by the borrowers. So, to tackle this problem, banks now a days use some models to predict the possibility of loan repayment from the borrower. Factors like annual income, employment status, home ownership, current debt etc are taken into consideration to categorize the loan request as bad loan or not. So, this paper basically aims to develop a similar model, but using ensemble machine learning algorithm of Random Forest Classification. And perform a comparative analysis with the model (Decision Tree Classification) that are currently in use. After complete implementation of all the models it was concluded that Random Forest Classifier Outperformed Decision Tree Classifier in terms of accuracy. © 2022 IEEE.