Data science and machine learning methods for detecting credit card fraud

Abstract

Credit card fraud remains a persistent challenge in the realm of financial security, necessitating innovative approaches for detection. This paper presents a comprehensive investigation into credit card fraud detection, focusing on integrating rule-based systems and machine learning methods to enhance accuracy and efficiency. The methodology encompasses data collection from a reputable source, thorough preprocessing, model development, and online execution. Performance evaluation employs a diverse array of metrics, including precision, recall, F1 score, accuracy, confusion matrix, false positive rate, learning curve, precision-recall curve, cumulative gains curve, and ROC curve. Results demonstrate a balanced trade-off between precision and recall, essential for effective fraud detection. Detailed discussions interpret these findings, offering valuable insights and avenues for future research. This research contributes to advancing fraud detection methodologies and holds promise for enhancing financial transaction security