A Review of Credit Card Fraud Detection Methods: An Approach Focused on Data and Technique

A crucial rule in the modern economy is the use of credit cards. It turns become an inevitable aspect of domestic, commercial, and international operations. Even while there are many advantages to using credit cards when done properly and cautiously, fraudulent activity can result in serious credit and financial harm. To address the rise in credit card fraud, numerous strategies have been put forth. Nevertheless, preventing credit card theft is the common objective of all these strategies; each has unique disadvantages, benefits, and traits. After looking into the challenges of detecting credit card fraud, we aim to cover the most recent methods, datasets, and assessment standards for credit card fraud detection in this paper. A list and comparison of the benefits and drawbacks of fraud detection techniques are provided. Additionally, the methodologies provided are categorised into two primary ways for detecting fraud: supervised misuse and unsupervised anomaly detection. Once more, the capacity to process both numerical and categorical datasets is used to classify the strategies. The useful and common characteristics of various datasets used in literature are then extracted for future use once they have been described and classified into real and synthesised data. Also gathered and addressed are evaluation criteria used in the literature. Thus, unresolved problems in detecting credit card fraud are described as recommendations for novice researchers.