A Comparative Analysis of Credit Card Fraud Detection Machine Learning Algorithms

Because of the development and quick expansion of E-Commerce, the usage of credit cards for online purchases has skyrocketed, resulting in an explosion in credit card fraud. As credit cards become the most prominent means of payment for both digital and in-store purchases, the number of incidents of fraud related to them rises. In practical instances, determining whether a transaction is legitimate or fraudulent is a complicated and demanding undertaking. As a result, all credit card issuing organizations must implement robust fraud detection systems to prevent losses. This study attempts to compare different strategies and determine which one is the most efficient. This study used Support Vector Machine, Logistic Regression, Random Forests, K Nearest Neighbors and gradient boosting frameworks including Xgboost, LightGBM, and CatBoost Models are built using the European Dataset. The outcomes are analyzed using three parameters: precision, recall, and F1 score. This research aids in determining the most efficient algorithm among the many algorithms employed.