MACHINE LEARNING FOR ROAMING FRAUD DETECTION USING TAP RECORDS

The proposed Machine Learning-based Roaming Fraud Detection System using TAP Records provides an intelligent approach for identifying suspicious roaming activities and reducing potential financial losses for mobile network operators. By analyzing subscriber behavior, visited-network information, service usage, transaction values, roaming duration, and other TAP-derived characteristics, the system can identify abnormal patterns that may indicate fraudulent activity. The methodology combines data preprocessing, behavioral feature extraction, feature selection, and supervised machine-learning classification to distinguish legitimate and fraudulent roaming transactions. Experimental comparison indicates that ensemble learning approaches, particularly XGBoost and Random Forest, can provide strong classification performance for complex telecom-fraud patterns. Previous research has also demonstrated the effectiveness of machine learning for mobile roaming-fraud detection, while graph-based telecom-fraud research shows the value of modeling subscriber behavior and relationships. The proposed framework can support automated fraud screening, risk scoring, and alert generation, thereby reducing dependence on manual investigation. Future integration with near-real-time roaming feeds such as NRTRDE can further reduce detection latency, while TAP records can support subsequent billing validation and settlement analysis. Overall, the proposed approach provides a scalable foundation for intelligent roaming-fraud management.

[1] S. M. Bhat, “A Review of Telecommunication Fraud Detection Techniques,” Zenodo, 2021. [2] K. Cherladine, “Innovative Telecom: MVNOs and MVNAs Leveraging Cloud-Based Technologies,” International Journal of Computer Trends and Technology, vol. 72, no. 8, pp. 190–198, 2024. [3] Y. Zhang, X. Wang, and Z. Li, “BTG: A Bridge to Graph Machine Learning in Telecommunications Fraud Detection,” Future Generation Computer Systems, vol. 137, pp. 274–287, 2022, doi: 10.1016/j.future.2022.07.020. [4] A. Ravi, M. Msahli, H. Qiu, G. Memmi, A. Bifet, and M. Qiu, “Wangiri Fraud: Pattern Analysis and Machine-Learning-Based Detection,” IEEE Internet of Things Journal, vol. 10, no. 8, pp. 6794–6802, 2023, doi: 10.1109/JIOT.2022.3174143. [5] K. Cherladine, “Harnessing Artificial Intelligence for Operational Excellence and Revenue Maximization in Mobile Virtual Network Operator (MVNO) Operations,” International Journal of Management, IT & Engineering, vol. 14, no. 5, pp. 32–35, 2024. [6] S. R. Pratihar, S. Paul, P. K. Dash, and A. K. Das, “Fraud Analytics Using Machine-Learning & Engineering on Big Data (FAME) for Telecom,” arXiv preprint, 2023. [7] “NFA: A Neural Factorization Autoencoder Based Online Telephony Fraud Detection,” Digital Communications and Networks, vol. 10, no. 1, pp. 158–167, 2024, doi: 10.1016/j.dcan.2023.03.002. [8] D. Chen and Y. Wu, “Research on the Use of Communication Big Data and AI Artificial Intelligence Technology to Construct Telecom Fraud Prevention Behavior Portrait,” Intelligent Decision Technologies, vol. 18, no. 3, 2024, doi: 10.3233/IDT-240386. [9] K. Cherladine, “Navigating Security and Privacy: Best Practices in MVNO Networks,” 2024/06, vol. 12, no. 6, pp. 1–4, 2024. [10] “Innovative Telecom Fraud Detection: A New Dataset and an Advanced Model with RoBERTa and Dual Loss Functions,” Applied Sciences, vol. 14, no. 24, Art. no. 11628, 2024, doi: 10.3390/app142411628.