A MACHINE LEARNING-BASED LOAN ELIGIBILITY PREDICTION USING BANKING DATA WITH CSV LOGGING

ABSTRACT In the evolving landscape of financial technology, the need for accurate, transparent, and efficient loan approval systems has become paramount. This project presents the development of a machine learning-based system for predicting loan eligibility using real-world banking data. Which addresses the limitations of traditional manual methods—such as bias, inconsistency, and lack of scalability—by leveraging supervised learning algorithms including Decision Trees, Random Forests, and Logistic Regression. The system is trained on a publicly available dataset containing essential financial indicators such as income level, employment status, credit history, and loan amount. A key innovation of this project is the integration of CSV logging, which ensures traceability, auditability, and transparency of all model decisions. Each prediction, along with its input parameters and timestamps, is recorded in a structured CSV format, enabling financial institutions to meet compliance standards and perform post-decision audits. The design follows a modular architecture encompassing data preprocessing, model training, user interface development, and logging subsystems. The project adopts the CRISP-DM methodology, ensuring a robust process from business understanding to deployment. Evaluation metrics such as accuracy, precision, recall, and F1-score were used to assess the performance of the models. The results demonstrate that the system achieves high predictive accuracy while maintaining accountability through detailed logging. This solution offers a scalable and interpretable framework for financial institutions seeking to modernize their loan approval processes with machine learning.