Machine Learning Approaches for Credit Card Fraud Detection: A Comparative Analysis and the Promise of 1D Convolutional Neural Networks

The increase in online financial transactions, reinforced by worldwide trends and the COVID-19 epidemic, has regrettably resulted in an unparalleled increase in credit card fraud. The banking sector has seen substantial financial losses as a result of payment fraud, amounting to an estimated USD 40.62 billion by 2027. Given this concerning pattern, there is an increasing demand for proactive and technology-driven techniques to detect fraud, as conventional rule-based systems are insufficient.This study explores the field of machine learning in the context of detecting credit card fraud. Machine learning, known for its capacity to facilitate computer learning without explicit programming, presents a possible alternative. It tackles obstacles including heavily skewed data distributions and the dynamic characteristics of fraudulent activity. This research takes an additional step by suggesting and assessing various hybrid models for the identical dataset. The study carefully examines six supervised machine learning algorithms: Support Vector Machine (SVM), Random Forest (RF), 1D-CNN, eXtreme Gradient Boosting (XGBOOST), Logistic Regression, and Adaptive Boosting (AdaBoost) on three different datasets. The performance of these hybrid models is thoroughly evaluated and compared to state-of-the-art machine learning methods using real- world data and a variety of assessment measures.The results indicate that Random Forest achieved outstanding performance, while 1D-CNN exhibited a gradual and consistent learning trajectory, suggesting potential for further investigation. However, in terms of performance, Random Forest outperforms 1D-CNN with low complexity and faster execution. This research enhances the area of credit card fraud detection by providing valuable information on the relative efficacy of different machine learning models and emphasizing the potential of deep learning methods in tackling this crucial problem.