DEVELOPMENT OF AN ENHANCED SMALL AND MEDIUM-SCALE ENTERPRISE LOAN DISTRIBUTION SYSTEM USING AN ENSEMBLE METHOD

Small and Medium-Scale Enterprises (SMEs) are critical to Nigeria's economic growth, yet many face persistent barriers to accessing timely and affordable financing.Traditional loan distribution systems often rely on manual, subjective assessments that are inefficient, biased, and limited in scope.This study presents the design and implementation of an Enhanced SME Loan Distribution System leveraging ensemble machine learning methods: Random Forest, XGBoost, and Logistic Regression to improve loan approval accuracy, efficiency, and fairness.Using a real-world SME loan dataset, the system applies data preprocessing, feature engineering, and model integration through a voting ensemble approach.Performance evaluation shows the ensemble model outperforms baseline classifiers, achieving 82.4% accuracy, 81.3% precision, 80.5% recall, and an ROC-AUC score of 0.86.The system also demonstrated robustness in varied data scenarios and improved decision-making transparency.Key contributions include a scalable framework for SME loan assessment, integration of multiple predictive models, and a user-friendly interface for lenders.This work advances the application of machine learning in financial decision-making and offers practical implications for enhancing financial inclusion in developing economies.Recommendations are provided for improving model generalizability, interpretability, and compliance with ethical lending practices.