Transforming Financial Lending: A Scalable Microservices Approach using AI and Spring Boot

Abstract

The financial services sector has seen substantial transformation in recent years due to the use of microservices architecture, artificial intelligence, and other modern frameworks. Loan default prediction and fraud detection problems in financial lending processes now receive solutions through the use of artificial intelligence (AI) and machine learning (ML) techniques, since lenders want more secure, scalable, and improved systems. The proposed method details the complete implementation of a financial lending system transformation through AI while supporting it with Spring Boot microservices infrastructure. The research utilizes the Lending Club Loan Dataset, containing more than 100,000 loan records, to apply ML models for loan default prediction while improving lending decision-making. The research design applies important preprocessing methods that use ANOVA feature selection alongside methods to impute missing values and eliminate outliers to guarantee data reliability. The model used Logistic Regression (LR) because of its straightforward nature as an effective tool for binary classification operations. The performance metrics of the LR model showed outstanding results, reaching 93% accuracy and equal precision and recall ratings at 93%, besides an F1-score value of 92%. The LR model provided better predictive capabilities than the competing alternatives, which included Boost, Decision Tree (DT), and Gradient Boosting. Microservices architecture combined with AI and ML demonstrates great promise in transforming financial lending operations because it enables better and faster decision-making as well as operational efficiency and scalability.