Authors: L. Rahul Chandra, Kekkarla Madhu, V. Shirisha, Atla Sonya
Abstract: The future of loan approvals is increasingly driven by Artificial Intelligence (AI), offering faster and data-informed decisions. However, traditional machine learning (ML) models often lack transparency, making them unsuitable for high-stakes financial decisions. This paper presents an Explainable AI (XAI) framework based on a Belief Rule Base (BRB) to automate and enhance the loan underwriting process. The BRB model combines expert knowledge with supervised learning and supports both factual and heuristic rules within a hierarchical structure. The system provides clear, interpretable explanations by highlighting activated rules and the influence of input attributes, ensuring transparency and regulatory compliance. A case study on mortgage underwriting demonstrates the model’s ability to balance accuracy with explainability, outperforming conventional black-box approaches in trust and interpretability. This work underscores the potential of XAI to shape a fairer, more transparent future for automated loan approvals.
DOI: http://doi.org/
