International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 09 | Sep 2025
p-ISSN: 2395-0072
www.irjet.net
A Comparative Study of Machine Learning Algorithms for Credit Card Fraud Detection Rifat Perween1, Nisha Kumari Singh2 12Department of Computer Science & Technology
Usha Mittal Institute of Technology, Shreemati Nathibai Damodar Thackersay Women’s University, Mumbai, India. ---------------------------------------------------------------------***--------------------------------------------------------------------1.1 Problem Statement Abstract - Credit card fraud refers to the physical loss of credit card or loss of sensitive credit card information. Many machine learning algorithms can be used for detection. This research shows several algorithms that can be used for classifying transactions as fraud or genuine ones. Credit Card Fraud Detection dataset was used in the research. Because the dataset was highly imbalanced, to solve the issue of class imbalance, we re-sampled the dataset using the Synthetic Minority over-sampling Technique (SMOTE). This framework was evaluated using several machine learning (ML) methods, including Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Decision Tree (DT), and Adaptive Boosting (AdaBoost). The models were evaluated using the accuracy, the recall, and the precision. The proposed model can be used for the detection of other irregularities.
With the exponential growth in digital transactions, credit card fraud has become a pressing concern for financial institutions worldwide. Fraudulent activities not only result in significant financial losses for both consumers and banks but also undermine trust in digital payment systems. Traditional fraud detection systems often rely on rule-based algorithms that can no longer cope with the dynamic and sophisticated nature of modern fraud. Therefore, there is a critical need for an advanced fraud detection system that leverages machine learning to identify and prevent fraudulent transactions in real time. The challenge of addressing the highly imbalanced nature of Credit Card Fraud Detection datasets. Firstly, the dynamic nature of fraudulent activities necessitates a model that is not only accurate but also effective. Fraudsters continuously innovate their strategies to evade detection, which means the model must be capable of learning from new data in real-time or near-real-time to stay effective. Selection of appropriate machine learning algorithms for credit card fraud detection. It’s important to identify the most suitable algorithm that demonstrates high accuracy. The Synthetic Minority Over-sampling Technique (SMOTE) is employed. This framework was evaluated with various models such as Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Decision Tree (DT), coupled with Adaptive Boosting (AdaBoost), to ensure high accuracy in detecting fraudulent transactions.
Key Words: Fraud detection, Applications of Machine Learning, XGBoost, Decision Tree, Random Forest
1. INTRODUCTION A credit card is typically issued to customers, enabling them to buy goods or services within a credit limit or withdraw cash in advance. It offers users the advantage of time, allowing them to repay later within a specified timeframe, often extending to the next billing cycle. Credit card frauds are susceptible and attractive targets. Perpetrators can withdraw a significant amount swiftly and without the owner’s awareness. The challenge in detecting fraud arises from fraudsters attempting to make their activities seem legitimate, adding complexity to the task of identifying fraudulent transactions. Web payment gateways have recently become popular for card-notpresent transactions in credit card operations. “The number of reports of identity theft climbed by 113 percent between 2019 and 2020, while the number of reports of identity theft using credit cards rose by 44.6 percent. Of the almost 1.4 million instances of identity theft in 2020, 393,207 involved credit card fraud. As a result, credit card fraud surpassed government documents and benefits fraud as the second most frequent identity theft recorded crime for the year. Machine learning is the solution for detecting the issues on large databases which are impossible for humans.
© 2025, IRJET
|
Impact Factor value: 8.315
2. LITERATURE SURVEY Several studies have explored machine learning and ensemble-based approaches for credit card fraud detection, addressing challenges such as class imbalance, evolving fraud patterns, and the need for real-time performance. Researchers have experimented with traditional classifiers, ensemble models, and advanced neural networks to improve detection accuracy and adaptability. Techniques like oversampling, feature engineering, and hybrid methods have also been employed to enhance model robustness. The following review discusses key contributions from existing literature and their findings.
|
ISO 9001:2008 Certified Journal
|
Page 455