Akademik Veri Yönetim Sistemi
Borsa Istanbul Review, 2026 (SSCI, Scopus)
This research explores the relationship between exposure to online payment fraud and Digital Financial Literacy (DFL) through Explainable Artificial Intelligence (XAI) frameworks. The rapid adoption of accessible digital payment methods has disproportionately heightened the risk of financial fraud for individuals exhibiting limited financial literacy or suboptimal digital behaviors. To quantify these risks, this study proposes a novel “Financial Negligence Score” (FNS) as a behavioral proxy to measure situational vulnerability, user attentiveness, and risk awareness during online financial transactions. Machine learning algorithms and statistical models—including Random Forest, XGBoost, CatBoost, and Logistic Regression—are deployed to evaluate the predictive capacity of these behavioral indicators. The empirical findings indicate that transactions characterized by financially inattentive behaviors exhibit a significantly higher probability of fraudulent outcomes. Furthermore, integrating the financial negligence score substantially improves the predictive performance and reliability of fraud detection models, particularly within ensemble learning architectures. Model interpretations derived via SHAP analysis confirm that behavioral traits serve as critical features in fraud classification. This study contributes to the existing literature by: (1) operationalizing an innovative, data-driven behavioral proxy for DFL; (2) leveraging XAI techniques to elucidate the underlying mechanics of fraud prediction; and (3) providing robust empirical evidence that elevated digital financial awareness directly correlates with a reduced likelihood of fraud victimization. These insights offer valuable implications for targeted financial education, dynamic fraud prevention strategies, and the design of safer digital financial ecosystems.