Explainable AI for Fraud Detection in Financial Transactions
Main Article Content
Abstract
The study hypothesizes the ideas of incorporating Explainable Artificial Intelligence (XAI) into financial fraud detection models to achieve compliance with regulatory demands on transparency. An ensemble-based model (e.g. XGBoost, LightGBM, and CatBoost) is integrated with XAI tools (e.g. SHAP, LIME, and PDP). The method has high performance and explains its outputs, where predictive power and stakeholder trust are guaranteed. The model gets an AUCROC of 0.99 on the IEEE-CIS Fraud Detection dataset. The research illustrates the fact that one can develop AI-based financial systems that are not only efficient in terms of fraud detection capabilities but also adherent to regulations focused on transparency.
Article Details
Issue
Section
Articles