Journal of Emerging Trends in Engineering and Applied Sciences: Research Areas

Article Title: AI-Powered Detection Of Cryptojacking On Android Devices: Bridging Global Cyber Threats And Local Realities In Emerging Economies
by Augustine Ndudi Egere, Callistus Tochukwu Ikwuazom,Grace Amina Onyeabor, Ngozi Ukamaka Okonkwo, Nwokocha Ihe Aaron, Ezeaku Francisca Ngozi

Abstract:
Cryptojacking, or the unauthorized use of mobile devices for cryptocurrency mining, poses a growing threat to cybersecurity in developing economies where the Android operating system plays a major role, and existing detection systems are too complex to be computed on mobile devices. This paper explores the possibility of using lightweight AI to detect cryptojacking attacks in low-resource developing countries on Android devices. A comprehensive data set of 8,500 Android applications (4,200 malicious and 4,300 benign) was compiled in three African countries, using threat intelligence feeds, honeypots, and verified malware repositories. Several machine learning and lightweight deep learning algorithms: Random Forest, XGBoost, CatBoost and quantized neural networks were systematically evaluated using repeated 5?fold cross?validation and real?world deployment. Random Forest gives the best results (F1-gain: 0.969), a small model size (3.2?MB), a minimal latency for inference (12.4?ms) and a low power usage impact (2.1%). The deployment in the field was able to be implemented with 2,400 participants with a result of an 87.7% reduction in infection rates when compared with the control group. Furthermore, explainable AI greatly enhanced user interaction, alert understanding, and reaction actions. The outcomes demonstrate that lightweight solutions can be effective in high detection rates in low resource areas, providing scalable and easy-to-use cybersecurity to support global cyber threats and local digital realities. __________________________________________________________________________________________ Keywords:

Keywords: Cryptojacking Detection, Android Malware, Machine Learning, Emerging Economies

Download full paper