A Lightweight ANN Model for Mobile Sim Swap Fraud Prevention in Resource Constrained Environments

SIM swap fraud has become a critical threat to the security and privacy of mobile phone users, particularly in resource-constrained environments. This type of fraud involves cybercriminals deceiving mobile network operators into transferring a victim's phone number to a new SIM card under their control. Once successful, the attackers can intercept messages and calls, including sensitive verification codes, allowing them to bypass two-factor authentication, reset passwords, and gain unauthorized access to financial and personal accounts. To address this issue, a lightweight Artificial Neural Network (ANN)-based fraud prevention and detection system was developed. The system is designed to detect abnormal patterns associated with SIM swap activities, such as irregular changes in network authentication and user location data. The ANN model was trained using supervised learning, where network behavior data was fed into the model, and weights were adjusted through backpropagation to optimize detection accuracy.The model distinguishes between normal user behavior and SIM swap anomalies, enabling real-time fraud detection. Upon detection of a suspected SIM swap, the system automatically initiates preventive actions, such as blocking unauthorized access and preserving user data integrity. The system architecture emphasizes minimal resource consumption, making it suitable for deployment in low-resource environments such as rural or developing regions.The Object-Oriented Analysis and Design Methodology (OOADM) was employed in modeling and implementing the system as a collection of interacting classes and objects. Python was used for system development. Validation by a mobile operator's fraud management unit confirmed the system’s effectiveness, with a detection accuracy of 95%. These results affirm the potential of lightweight ANN models in enhancing SIM swap fraud prevention in environments with limited computational resources.