Machine Learning Models for Predicting Bank Loan Eligibility

Abstract

Machine learning algorithms are revolutionizing processes in all fields including; real-estate, security, bioinformatics, and the financial industry. The loan approval process is one of the most tedious task in the banking industry. Modern technology such as machine learning models can improve the speed, efficacy, and accuracy of loan approval processes. This paper presents six (6) machine learning algorithms (Random Forest, Gradient Boost, Decision Tree, Support Vector Machine, K-Nearest Neighbor, and Logistic Regression) for predicting loan eligibility. The models were trained on the historical dataset ‘Loan Eligible Dataset,’ available on Kaggle and licensed under Database Contents License (DbCL) v1.0. The dataset was processed and analyzed using Python programming libraries on Kaggle’s Jupyter Notebook cloud environment. Our research result showed high-performance accuracy, with the Random forest algorithm having the highest score of 95.55% and Logistic regression with the lowest score of 80%. Our Models outperformed two of the three loan prediction models found in the literature in terms of precision-recall and accuracy.