The lack of a robust framework that addresses temporal drift in credit card fraud detection

The gap

Evidence profile

Sourced from the future work and recommendations and future-work section and stated research gap of the source papers, classified as general, drawn from work published between 2025 and 2026, spanning 6 journals. Those papers have been cited 19 times in total.

Supporting evidence — 6 representative gaps

• Online Payment Fraud Detection Using Machine Learning (2026) · International Journal of Latest Technology in Engineering Management & Applied Science · doi

  In this work, an effective online payment fraud detection system was developed using machine learning techniques. The combination of CatBoost and XGBoost models through an ensemble approach resulted in improved predictive performance. PCA was used for dimensionality reduction, and SMOTE was applied to address class imbalance, which significantly enhanced the model’s ability to detect fraudulent transactions. The system achieved high accuracy, precision, recall, and AUC scores, demonstrating its effectiveness in real- world scenarios. Additionally, the deployment of the model using Streamlit provides a practical interface for real-time fraud detection. Future work can focus on integrating deep learning approaches such as neural networks and graph-based models to capture more complex transaction patterns. Furthermore, real-time streaming data and largescale deployment can be explored to improve scalability and adaptability in dynamic financial environments. REFERENCES 1. M. Habibpour, H. Gharoun, M. Mehdipour, A. Tajally, H. Asgharnezhad, A. Shamsi, A. Khosravi, M. Shafie-Khah, S. Nahavandi, and J. P. S. Catalao, ''Uncertainty-aware Online payment fraud detection using deep learning 2021; arXiv:2107.13508. 2. A. Cherif, A. Badhib, H. Ammar, S. Alshehri, M. Kalkatawi, and A. Imine. "Online payment fraud detection in the era of disruptive technologies: A systematic review." J. King Saud Univ. Computer and Information Science, vol. 35, no. 1, pp. 145-174, Jan. 2023, doi:10.1016/j.jksuci.2022.11.008. 3. T. K. Dang, T. C. Tran, L. M. Tuan, and M. V. Tiep. "Machine learning based on resampling approaches and deep reinforcement learning for Online payment fraud detection systems." Appl. Sci., vol. 11, no. 21, p. 10004, Oct. 2021; doi: 10.3390/app112110004. Page 927 www.rsisinternational.org INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING, MANAGEMENT & APPLIED SCIENCE (IJLTEMAS) ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue IV, April 2026 4. Chaquet-Ulldemolins et al., ''On the black-box problem for fraud detection using machine learning (I): Linear models and informative feature selection,'' Applied Sciences, vol. 12, no. 7, p. 3328, March 2022, doi: 10.3390/app12073328. 5. E. F. Malik, K. W. Khaw, B. Belaton, W. P. Wong, and X. Chew. "Online payment fraud detection using a new hybrid machine learning architecture." Mathematics, vol. 10, no. 9, p. 1480, April 2022; doi: 10.3390/math10091480. 6. I. Benchaji, S. Douzi, B. El Ouahidi, and J. Jaafari, "Enhanced Online payment fraud detection using attention mechanism and LSTM deep model," J. Big Data, vol. 8, no. 1, p. 151, December 2021; doi: 7. 10.1186/s40537-021-00541-8. 8. E. Esenogho, I. D. Mienye, T. G. Swart, K. Aruleba, and G. Obaido. "A neural network ensemble with feature engineering for

  generalfuture work
  Keywords: fraud detection learning online payment using machine deep models applied model real system ensemble enhanced
• Bidirectional fusion heterogeneous graph networks for semi-supervised Bitcoin transaction anomaly detection in dynamic transaction graphs (2026) · PLOS One · doi

  5.1 Conclusion In this paper, we carry out systematic research on bitcoin transaction anomaly detection task and propose several inno- vative methods: firstly, to meet the practical requirements, we define the dynamic heterogeneous graph semi-supervised bitcoin anomaly detection task and design the Bi-directional Fusion Heterogeneous Graph Network (BF-HGN) to construct the basic framework. Second, in feature extraction, we improve upon RGCN to construct EvolveRGCN and combines EvolveGCN to design a gradual scheme. It also introduces LSTM to capture temporal features and deeply mines dynamic features through a fusion strategy. Further, we propose the Multi-type Feature Fusion Extractor. This improves the dynamic relationship modeling capability by capturing the upper and lower time-point subgraph associations. Lastly, we address the class imbalance problem caused by unlabeled anomalous samples by designing Class-balanced Classifiers. These classifiers balance the training data class distribution by generating pseudo-abnormal nodes constrained by AA and AFSR loss function. 5.2 Outlook Future research can be extended to a broader range of financial transaction scenarios, thereby strengthening risk pre- vention and control capabilities. Further exploration of the optimization space of feature extraction and fusion strate- gies reveals potential associations in complex data and injects richer semantic information into the model. Meanwhile, continuous efforts should be made to refine the optimization path of loss functions to improve the generation quality of pseudo-anomalous nodes, so as to promote the security and stability of anomaly detection technologies in Bitcoin transactions and related fields. In addition to technical advancements, future studies should incorporate regulatory, ethical, and societal considerations into the design of anomaly detection systems. Inspired by the sociotechnical framework proposed by Rahman et al. [62], responsible and trustworthy FinTech development can be better supported in blockchain transaction surveillance, particularly with respect to regulatory compliance, transparency, and social accountability.

  generalfuture work
  Keywords: anomaly detection fusion bitcoin transaction dynamic design feature class task propose heterogeneous graph construct framework
• Integrating Support Vector Machine Classifiers for Real-Time Sentiment Analysis and Fraud Detection in A Fashion E-Commerce Platform (2026) · Scientific Journal of Engineering, and Technology · doi

  Future applications of similar systems should incorporate a real payment gateway instead of simulated transaction data from the beginning of the project. Training datasets need to be continually increased and updated to ensure that the performance of the model remains accurate as fraud and review language patterns change over time. Developers need to make sure that their pipelines are continuous for retraining, not just one-time deployments. Any company that plans to implement a similar design should test extensively and in a staging environment before deployment, especially the fraud detection element, which seems to have operational costs both from false positives and false negatives. REFERENCES Banu, R., Ashok, A., Dwivedi, V. K., Reddy, K. A., Thulasimani, T., & Nishant, N. (2024). An innovative method for fraud detection in e-commerce using DCNN-multiclass SVM model. In 2024 International Conference on Intelligent Algorithms for Computational Intelligence Systems (IACIS) (pp. 1-6). IEEE. https://doi.org/10.1109/IACIS61494.2024.10721774 Coherent Market Insights. (2023). Fashion e-commerce market analysis and growth projections. Retrieved from https:// www.coherentmarketinsights.com Kaggle. (2023a). Credit card fraud detection dataset. Retrieved https://www.kaggle.com/datasets/mlg-ulb/ from creditcardfraud Kaggle. (2023b). Women's e-commerce clothing reviews dataset. Retrieved from https://www.kaggle.com/datasets/ nicapotato/womens-ecommerce-clothing-reviews Kumar, S., Gunjan, V. K., Ansari, M. D., & Pathak, R. (2022). Credit card fraud detection using support vector machine. In Proceedings of the 2nd International Conference on Recent Trends in Machine Learning, IoT, Smart Cities and Applications: ICMISC 2021 (pp. 27-37). Springer, Singapore. https://doi.org/10.1007/978-981-16-6407-6_3 Mutemi, A., & Bacao, F. (2024). E-commerce fraud detection based on machine learning techniques: Systematic literature review. Big Data Mining and Analytics, 7(2), 419-444. https:// doi.org/10.26599/BDMA.2023.9020024 Sharma, H. D., & Goyal, P. (2024). Interpretable aspect based sentiment classification of online educational reviews using SVM model and explainable LIME-AI model. International Journal of Information Technology, 16, 4567-4578. https://doi. org/10.1007/s41870-024-02125-w Shopify. (2024). E-commerce statistics and trends: Social commerce revenue data. Retrieved from https://www. shopify.com/research Statista. (2023). Global fashion e-commerce market size and forecasts 2024-2030. Retrieved from https://www.statista.com Tabany, M., & Gueffal, M. (2024). Sentiment analysis and fake Amazon reviews classification using SVM supervised machine learning model. Journal of Advances in Information Technology, 15(1), 49-58. https://journals.stecab.comStecab Publ

  generalrecommendations
  Keywords: https commerce fraud model detection retrieved using kaggle reviews machine datasets international market learning applications
• Enhancing credit card fraud detection using traditional and deep learning models with class imbalance mitigation (2025) · Frontiers in Artificial Intelligence · cited 19× · doi

  This study investigated the effectiveness of various machine learning approaches; Logistic Regression, Decision Tree, Random Forest, and an Enhanced Deep Learning model for the detection of fraudulent credit card transactions. To address the severe class imbalance inherent in the dataset, the Synthetic Minority Over- sampling Technique (SMOTE) was employed, resulting in significant performance improvements across all models. Among the traditional models, Random Forest achieved the highest overall performance with an accuracy of 99.95%, an F1 score of 0.8256, and a ROC-AUC of 0.9759. The Deep Learning model, enhanced with focal loss and regularization techniques, demonstrated the highest precision and a competitive F1 score, indicating its ability to reduce false positives while maintaining high recall. These results affirm that combining advanced sampling methods like SMOTE with both classical and deep learning models substantially improves fraud detection accuracy and reliability. Moreover, the enhanced deep learning model’s stability during training and strong generalization performance underscores its suitability for complex fraud detection tasks. Future work should focus on expanding detection capabilities beyond isolated transactions to uncover fraud rings, which involve coordinated fraudulent activities across multiple accounts. Graph- based learning methods, particularly graph neural networks (GNNs), offer strong potential for capturing such relational dependencies. Furthermore, the development of federated learning frameworks can enable collaborative fraud detection across institutions while preserving data privacy, a critical requirement in financial applications. Another promising direction is the integration of AI with blockchain technologies to enhance transparency, traceability, and auditability of financial transactions, as highlighted in recent reviews (e.g., Ressi et al., 2024). Finally, validating the proposed models across multiple benchmark datasets

  generalfuture work
  Keywords: learning detection deep across models fraud enhanced model transactions performance random forest fraudulent sampling smote
• Intelligent Credit Card Fraud Detection (2026) · Zenodo (CERN European Organization for Nuclear Research) · doi

  Future research should focus on developing more advanced machine learning and hybrid methods for fraud detection. The use of graph-based approaches and deep learning models, such as Convolutional Neural Networks (CNNs) and Graph Neural Networks (GNNs), should be explored. The development of more effective methods for handling imbalanced datasets is a future research direction.

  generalfuture-work sectionevidence 5/5
  Keywords: future research focus developing advanced machine learning hybrid
• A robust machine learning framework for detecting temporal drift in financial fraud prevention (2026) · Scientific Reports · doi

  The lack of a robust framework that addresses temporal drift in credit card fraud detection. The need for a framework that combines feature engineering, oversampling, and ensemble methods to improve detection performance. The requirement for a framework that is evaluated on a benchmark dataset and compared to existing techniques.

  generalstated research gapevidence 5/5
  Keywords: lack robust framework addresses temporal drift credit card

Questions about this gap