Credit Card Fraud Detection Using Machine Learning: A Comparative Study of Ensemble Learning Algorithms

The enormous increase in credit card usage for both online and offline transactions have resulted in a similar increase in fraudulent activity, elevating credit card fraud to a serious worry for both financial institutions and consumers. This research endeavour entails an examination of conventional machine learning (ML) techniques, including decision tree, logistic regression, support vector machine, and Naive-Bayes, in conjunction with ensemble learning methodologies, such as XGboost, random forest, voting, gradient boosting, Adaboosting, and stacking, for the purpose of identifying fraudulent credit card transactions. The findings indicate that the random forest and XGBoost algorithms exhibited remarkable efficacy, attaining notable degrees of precision and F1-score. The study highlights the importance of utilising anomaly detection techniques in identifying fraudulent transactions. Moreover, it showcases the effectiveness of ML algorithms in augmenting the security of online transactions. Financial institutions can mitigate customer financial losses by employing real-time fraud detection through sophisticated ML methodologies. The research provides significant perspectives on the effectiveness of diverse ML algorithms in identifying credit card fraud and highlights the potential for further progress in this field.