The datasets included in this research are sourced from the Kaggle platform, specifically the IEEE-CIS Fraud Detection, PaySim Mobile Money Simulation, and BankSim Synthetic Transaction Data. Each dataset comprises transactional records that include temporal, numerical, and categorical elements, such as transaction amount, type, origin and destination accounts, device information, and time features. The IEEE-CIS dataset includes e-commerce and card transactions, whereas PaySim and BankSim simulate mobile and banking contexts. These datasets demonstrate significant class imbalance, with fraudulent instances comprising fewer than 1% of the total samples. Let 𝑁 and 𝑁 represent fraudulent and genuine transactions correspondingly, resulting in an imbalance ratio