International Journal of Scientific Research and Engineering Development

The rapid growth of digital banking and online payment systems has significantly increased the risk of financial fraud. Modern financial institutions process millions of transactions daily, making it difficult to manually detect suspicious activities. Traditional rule-based fraud detection systems fail to adapt to evolving fraud patterns and often result in high false positives. Machine learning techniques provide an effective solution by analysing large volumes of transaction data and identifying complex behavioural patterns. Models such as Logistic Regression, Random Forest, and Gradient Boosting are widely used for fraud detection tasks. However, many of these models operate as lack boxes, limiting transparency and trust in financial decision-making. To address this issue, this research proposes an Explainable AI-driven fraud detection system that integrates machine learning models with SHAP (SHapley Additive Explanations). The system predicts fraudulent transactions while providing interpretable insights into feature contributions. Experimental results show that ensemble models such as Random Forest outperform traditional models, achieving high ROC-AUC scores. SHAP analysis highlights key factors such as transaction amount and fraud history as major contributors. The proposed system improves accuracy, transparency, and decisionmaking in financial fraud detection

Pranav Babar, Satish Gujar,"Explainable AI Driven Banking Fraud Detection System Using Machine Learning and SHAP with Power BI Analytics" International Journal of Scientific Research and Engineering Development, V9(2): Page(2299-2306) Mar-Apr 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.