Machine Learning Based Approaches to Detect Loan Defaulters

Consumers acquire many loans from banks when they need money, and banks provide many low-interest rates offers to entice people to take out loans. However, if consumers do not pay their loans on time, the bank may incur a significant loss. The problem statement seeks to categorize whether people can repay their debt, preventing banks from incurring substantial losses. Defaulters can bankrupt banks due to large loan non-payment, resulting in a financial crisis in the country or for any bank that provides the loan. Before issuing a loan to a person, a comprehensive check is performed on their profile to ensure that they do not default, but it is still difficult to determine who will default and who will not. Because the number of individuals taking out loans is increasing year after year, a system to identify and handle this rising problem is urgently needed to find a solution. As the number of people taking out loans increases, so will the number of defaulters. There are a variety of classification machine learning techniques and deep learning approaches that may be used to solve the difficulties. The study's primary goal is to compare and contrast the Random Forest, Logistic Regression, and XGBoost models to see which one performs and provides the best accuracy.