Published Paper Details:

Small and medium-sized businesses (SMEs) are crucial to the expansion of the economy and the creation of jobs, but they frequently struggle to get loans because of their inadequate credit history, lack of collateral, and incomplete financial documents. Many SMEs are financially opaque in the financing ecosystem since traditional credit scoring systems mostly rely on formal financial statements. In order to enhance SME loan default prediction in data-sparse settings, this study suggests an explainable multi-source machine learning architecture that incorporates alternative data sources such GST filings, digital payment behavior, cash-flow patterns, and transaction activities. To guarantee both forecast accuracy and transparency in lending decisions, the suggested system integrates Explainable Artificial Intelligence (XAI) methodologies with machine learning algorithms. While SHAP and LIME algorithms offer comprehensible explanations for loan approvals and rejections, models like Random Forest and XGBoost are employed for default prediction. The goal of the project is to develop a credit evaluation system that is transparent, dependable, and compliant with regulations in order to improve SME financial inclusion while upholding efficient risk management standards.