Credit Card Fraud Detection: A Comparative Study of Machine Learning and Deep Learning Methods

Credit card fraud has become a significant concern in the digital era, driven by the rise in online transactions and the sophistication of fraudulent activities. Traditional fraud detection systems are increasingly inadequate due to their static nature and limited adaptability to new attack patterns. In response, this study presents a comparative analysis of recent machine learning (ML) and deep learning (DL) techniques used for credit card fraud detection (CCFD). A total of 29 peer-reviewed studies published between 2019 and 2024 were reviewed, covering a range of ML models such as Decision Trees, Random Forest, XGBoost, and ensemble methods, alongside DL models including CNNs, LSTMs, AutoEncoders, and Graph Neural Networks. The analysis focuses on performance metrics, dataset characteristics, model limitations, and the effectiveness of imbalance handling strategies. Findings reveal that while DL models often achieve higher accuracy, they demand more computational resources, whereas ML models offer better efficiency and interpretability. The study concludes with a discussion on key challenges and suggests future research directions, including hybrid model development, improved imbalance handling, and real-time system deployment.