Machine Learning Based Loan Approval System: A Comparative Analysis

Loans are a major source of income in the banking industry so they are the main cause of profitability. Still, finding trustworthy applicants who would pay back their loans is difficult even with the great amount of loan applications. The banking sector depends more and more on huge amounts of data as digitization grows in order to make wise judgements. This paper suggests a mutual system based on predictive analysis to rank consumer wait times and project future results. Banks assess loan requests depending on applicant eligibility and needs as new loan application methods keep developing. Loan approval is decided upon by means of historical customer data analysis using machine learning algorithms. With results noted for future usage, the Random Forest (RF) Algorithm is specifically employed to precisely estimate a customer's likelihood of loan repayment. The proposed method lowers time complexity in addition to improving accuracy. Random Forest exceeded other models like Decision Tree (95.78%), Naive Bayes (94.38%), Support Vector Machine (SVM) (92.27%), Logistic Regression (90.40%), and K-Nearest Neighbours (59.02%). With the best accuracy at 96.49%. The most efficient method for loan approval system optimization is to be found by this work.