Prediction of bank Loan Status using Machine Learning Algorithms

Abstract

Major income of banks and any financial organization is generated by loans. Banks can issue loans only to specific authentic people or organizations due to restricted resources or credits. Those who actually can able to repay the taken loan amount along with interest are safe people to whom loan can be sanctioned, but finding eligible (safe) people is a monotonous process. The problem is addressed by various researchers in the literature, however, accuracy level of their models proposed is utmost of 80%. Hence in our work, we proposed a model in which various machine learning algorithms are aggregated with ensemble algorithms like bagging and voting classifiers. The pre-eminent objective of our work is to predict whether a particular person is eligible for the loan or not. Our proposed model reduces human efforts and processing time as well and produces more accurate results than existing models. Experimental results show that our model improves the performance of the existing model from 80% to 94%.

Performance Analysis of Ensemble Learning and Feature Selection Methods in Loan Approval Prediction at Banks

“…Initially, when the four algorithms were run separately, they gave an accuracy result of 82%. By feeding the results from the bagged classifiers to the voting classifier, the accuracy increased to 94% [15].…”