Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{207848,
        author = {Ms. Arya M. Palve and Prof. Jayaprabha M. Kanase and Prof. Dr. B. D. Phulpagar},
        title = {EXPLAINABLE FINANCIAL FRAUD DETECTION USING LIGHTGBM AND SHAP ANALYSIS: A COMPARATIVE STUDY WITH XGB-GP FRAMEWORK},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {3},
        pages = {2959-2968},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=207848},
        abstract = {Financial fraud detection has become increasingly important as the rapid growth of digital transactions exposes financial institutions to sophisticated and evolving fraudulent activities. Conventional machine learning models often achieve high predictive performance but provide limited interpretability, making them less suitable for high-stakes financial decision-making where regulatory compliance and stakeholder trust are essential. This study proposes an explainable financial fraud detection framework that integrates the Light Gradient Boosting Machine (LightGBM) algorithm with SHapley Additive exPlanations (SHAP) to simultaneously achieve robust classification performance and transparent model interpretation. The proposed framework is comparatively evaluated against an Extreme Gradient Boosting–Genetic Programming (XGB-GP) framework to assess both predictive effectiveness and explainability. The experimental workflow includes data preprocessing, handling class imbalance, feature engineering, hyperparameter optimization, and stratified cross-validation using a publicly available financial transaction dataset. Performance is assessed using accuracy, precision, recall, F1-score, area under the receiver operating characteristic curve (AUC-ROC), and computational efficiency. SHAP analysis is employed to quantify feature contributions, provide global and local explanations, and identify the most influential variables affecting fraud predictions. Comparative analysis demonstrates that the LightGBM-SHAP framework achieves competitive or superior detection capability while significantly improving model transparency and reducing inference time compared with the XGB-GP approach. The explainability component enables analysts to understand individual prediction behavior, thereby supporting regulatory requirements and enhancing confidence in automated fraud detection systems. The proposed framework offers a practical and scalable solution for real-time financial fraud detection by combining high predictive accuracy with interpretable artificial intelligence, making it suitable for deployment in modern financial environments where both performance and transparency are critical.},
        keywords = {Financial Fraud Detection, Explainable Artificial Intelligence (XAI), LightGBM, SHAP, XGBoost, Genetic Programming, Machine Learning, Class Imbalance, Financial Analytics.},
        month = {August},
        }