Abstract: Nowadays, credit cards are the most common method of payment both offline and online purchases due to advancements in electronic commerce and communication technology. Thusly, the gamble of misrepresentation related with these exchanges has expanded. Each year, fraudulent credit card activities lead to significant losses for businesses finances and individuals, with fraudsters continually devising new schemes. Detecting credit card theft remains a difficult task for researchers because of the complexity and creativity of fraudsters. The imbalance in datasets used for fraud detection algorithms further complicates this task. Therefore, There are pressing need for efficient and effective methods to identify fraudulent credit card transactions. This paper makes a new approach to tackle this issue: the Gradient Boosting Classifier, a machine learning tool. Experimental results, demonstrating 100% training accuracy and 91% test accuracy, indicate that Other machine learning methods are inferior to the proposed method techniques. 

Keywords: innovative approach, Gradient Boosting, machine learning

| DOI: 10.17148/IARJSET.2024.11758