Modeling Imbalanced Financial Transactions for Credit Card Fraud Detection Using Machine Learning and Deep Learning Techniques

Abstract

As the growth of electronic commerce and digital payment systems is increasing at a rapid pace, the menace of credit card fraud has surfaced as a highly advanced global threat with a huge financial loss of billions of dollars on a yearly basis. The conventional fraud detection systems using traditional rule-based systems or conventional machine learning techniques have become inadequate in dealing with the huge volume, velocity, and complex nonlinear characteristics of the transactions. This review aims at presenting a comprehensive review of the advanced techniques in the field of credit card fraud detection with special emphasis on the evolution of advanced deep learning techniques from conventional machine learning techniques in the time frame of 2023-2025. It highlights a critical review of conventional machine learning techniques such as Random Forest and Support Vector Machines with advanced deep learning techniques such as Convolutional Neural Network and Long Short-Term Memory networks. This demonstrates that the results of these individual models face difficulties in handling issues like class imbalance and the need for intensive feature engineering. On the other hand, recent hybrid models, such as the Deep Hybrid CLST model, have shown promising results in handling the problem of detecting financial fraud by integrating spatial feature learning and temporal sequence learning. Furthermore, the review has shown the increasing trend of data silos in the industry due to the enforcement of privacy policies and has introduced Federated Learning and Graph Neural Networks as possible solutions for collaborative learning in the detection of financial fraud. Lastly, the review has shown the difficulties in handling the problem of detecting financial fraud, including the interpretability of the model and scalability in real-time, and has shown the possible future direction of developing stronger and more efficient financial fraud detection systems by focusing on Explainable AI and distributed computing frameworks.

Introduction

The rapid growth of internet usage, especially during the COVID-19 pandemic, has increased dependence on online services and digital payments. As online transactions have increased, cyberattacks, phishing, data breaches, and credit card fraud have also become major security concerns. The increasing financial losses caused by digital fraud highlight the need for advanced and reliable fraud-detection systems.

Traditional methods such as Decision Trees, Random Forest, Support Vector Machines (SVM), and Logistic Regression have been widely used for fraud detection. However, these methods face important limitations, particularly the highly imbalanced nature of financial datasets, where fraudulent transactions usually represent less than 0.5% of all transactions. Traditional models may therefore achieve high overall accuracy while failing to identify fraudulent cases effectively.

To address these limitations, researchers have introduced techniques such as resampling, cost-sensitive learning, genetic algorithms for feature selection, and distributed computing frameworks such as PySpark. These approaches aim to improve detection accuracy, reduce computational costs, and make fraud detection more suitable for large-scale, real-time applications.

More recently, deep learning has become increasingly important. CNNs can automatically learn complex transaction features, while LSTM and RNN models are useful for analyzing the sequential behavior of transactions. Hybrid models combining CNNs with LSTMs and ensemble approaches have also been explored to improve fraud-detection performance.

Another important development is the use of Graph Neural Networks (GNNs). Instead of examining transactions independently, GNNs model relationships among cardholders, merchants, devices, and locations. This allows them to identify complex patterns such as coordinated fraud and fraud rings that may be difficult for traditional models to detect.

The text also identifies several major challenges in fraud detection:

• Class imbalance: Fraudulent transactions are rare, making detection difficult and making accuracy alone an unreliable evaluation measure.

• Real-time scalability: Advanced models can require significant computational resources when processing large numbers of transactions.

• Concept drift: Fraudsters continually change their methods, meaning detection models must adapt to evolving patterns.

• Feature selection: Manual feature engineering can be time-consuming and may overlook important patterns.

• False positives: Incorrectly blocking legitimate transactions can negatively affect customers.

• Cold-start problems: New customers and merchants may have insufficient historical data.

• Privacy and data sharing: Financial institutions cannot freely share sensitive customer information.

• Explainability: Complex deep-learning and graph-based models can be difficult to interpret.

Finally, the text highlights Federated Learning (FL) as an emerging solution for privacy-preserving collaborative fraud detection. Frameworks such as FinGraphFL combine federated learning with GNNs so that institutions can collaboratively train models without directly sharing sensitive customer data.

Conclusion

This review demonstrates the evolution of the detection of credit card fraud using machine learning techniques. Previ-ously, statistical approaches and basic machine learning tech-niques were used. Although these techniques are still valid in some cases, especially when interpretability and efficiency are key factors in the approach. However, more recent approaches have focused on the use of deep learning and the combination of machine learning techniques in the detection of fraud. The combination of these techniques has shown a significant improvement in the performance of the systems. Despite the improvements in the detection of fraud using machine learning techniques, there is a gap between the results of the

References

[1] S. P. Ripa, F. Islam, and M. Arifuzzaman, “The emergence threat of phishing attack and the detection techniques using machine learning models,” in 2021 International Conference on Automa-tion, Control and Mechatronics for Industry 4.0 (ACMI), Insti-tute of Electrical and Electronics Engineers Inc., Jul. 2021, doi: 10.1109/ACMI53878.2021.9528204. [2] R. J. Bolton and D. J. Hand, “Statistical Fraud Detection: A Review,” 2002. [3] E. W. T. Ngai, Y. Hu, Y. H. Wong, Y. Chen, and X. Sun, “The application of data mining techniques in financial fraud detection: A classification framework and an academic review of literature,” Decision Support Sys-tems, vol. 50, no. 3, pp. 559–569, 2011, doi: 10.1016/j.dss.2010.08.006. [4] “A Comprehensive Survey of Data Mining-based Fraud Detection Re-search.” [5] R. Qi, “An Empirical Study on Credit Risk Assessment Using Machine Learning: Evidence from the Kaggle Credit Card Fraud Detection Dataset,” Journal of Computer, Signal, and System Research, vol. 2, no. 5, p. 48, 2025, doi: 10.71222/7mn4wp34. [6] B. Mehar Sri Satya Teja, B. Munendra, and S. Gokulkrishnan, “A Research Paper on Credit Card Fraud Detection,” International Research Journal of Engineering and Technology, 2022. [7] E. Ileberi, Y. Sun, and Z. Wang, “A machine learning based credit card fraud detection using the GA algorithm for feature selection,” Journal of Big Data, vol. 9, no. 1, Dec. 2022, doi: 10.1186/s40537-022-00573-8. [8] L. Theodorakopoulos, A. Theodoropoulou, F. Zakka, and C. Halkiopou-los, “Credit Card Fraud Detection with Machine Learning and Big Data Analytics: A PySpark Framework Implementation,” Jul. 2024, doi: 10.20944/preprints202407.0022.v1. [9] R. Jayalakshmi, R. G. Suresh Kumar, and T. Thanushree, “A Survey on Credit Card Fraud Detection using Deep Learning Model,” International Research Journal on Advanced Engineering and Management, vol. 3, no. 4, pp. 1325–1334, Apr. 2025, doi: 10.47392/irjaem.2025.0216. [10] S. Makki, Z. Assaghir, Y. Taher, R. Haque, M. S. Hacid, and H. Zeineddine, “An Experimental Study With Imbalanced Classification Approaches for Credit Card Fraud Detection,” IEEE Access, vol. 7, pp. 93010–93022, 2019, doi: 10.1109/ACCESS.2019.2927266. [11] E. Ileberi, Y. Sun, and Z. Wang, “A machine learning based credit card fraud detection using the GA algorithm for feature selection,” Journal of Big Data, vol. 9, no. 1, Dec. 2022, doi: 10.1186/s40537-022-00573-8. [12] J. Jurgovsky et al., “Sequence classification for credit-card fraud de-tection,” Expert Systems with Applications, vol. 100, pp. 234–245, Jun. 2018, doi: 10.1016/j.eswa.2018.01.037. [13] M. Jabeen, S. Ramzan, A. Raza, N. L. Fitriyani, M. Syafrudin, and S. W. Lee, “Enhanced Credit Card Fraud Detection Using Deep Hy-brid CLST Model,” Mathematics, vol. 13, no. 12, Jun. 2025, doi: 10.3390/math13121950. [14] I. D. Mienye and Y. Sun, “A Deep Learning Ensemble With Data Resampling for Credit Card Fraud Detection,” IEEE Access, vol. 11, pp. 30628–30638, 2023, doi: 10.1109/ACCESS.2023.3262020. [15] Y. Chen, C. Zhao, Y. Xu, C. Nie, and Y. Zhang, “Year-over-Year Developments in Financial Fraud Detection via Deep Learning: A Systematic Literature Review,” Jul. 2025. [16] A. Cherif, H. Ammar, M. Kalkatawi, S. Alshehri, and A. Imine, “Encoder–decoder graph neural network for credit card fraud detection,” Journal of King Saud University - Computer and Information Sciences, vol. 36, no. 3, Mar. 2024, doi: 10.1016/j.jksuci.2024.102003. [17] Z. Xia and S. C. Saha, “FinGraphFL: Financial Graph-Based Federated Learning for Enhanced Credit Card Fraud Detection,” Mathematics, vol. 13, no. 9, May 2025, doi: 10.3390/math13091396. [18] Y. Chen, K. Zhang, H. Zhu, and Z. Qiu, “A Novel Federated Transfer Learning Framework for Credit Card Fraud Detection Under Hetero-geneous Data Conditions,” Risks, vol. 13, no. 11, Nov. 2025, doi: 10.3390/risks13110208. [19] M. T. Singh, R. K. Prasad, G. R. Michael, N. K. Kaphungkui, and N. H. Singh, “Heterogeneous Graph Auto-Encoder for Credit Card Fraud Detection,” Oct. 2024. [20] I. D. Mienye and N. Jere, “Deep Learning for Credit Card Fraud Detec-tion: A Review of Algorithms, Challenges, and Solutions,” IEEE Access, vol. 12, pp. 96893–96910, 2024, doi: 10.1109/ACCESS.2024.3426955.

Copyright

Copyright © 2026 Mansi Sharma, Amandeep Verma, Rajni Sobti. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.