A Novel Automated Information Retrieval on Credit Score for Bank Loan Approval Using Financial and Repayment Data Transactions by Comparing KNN Over Linear Regression

The study aims to refine bank loan approval processes by assessing borrower risk, thereby improving customer retention. We compare the efficacy of two machine learning algorithms: K-Nearest Neighbors (KNN) and Linear Regression, in categorizing loan applications as either acceptable or unacceptable risks. Utilizing a dataset, the algorithms were evaluated for predictive accuracy with an 80% power (G*Power), a 5% level of significance, and a 95% confidence interval using the mean and standard deviation as parameters. The findings reveal that Linear Regression achieved a higher predictive accuracy (74.44%) in comparison to KNN (69.83%), with a significant difference in performance (<0.05p < 0.05). Therefore, Linear Regression is more adept at predicting bank loan approval outcomes than KNN.