مسار بولونيا
EN

فاطمه رحيم ناصر علي

السنة عنوان البحث نشر البحث
2026 Optimized intrusion detection in loT networks: A machine learning and genetic algorithm approach AIP Conference Proceedings
Abstract. The Internet of Things (IoT) is growing rapidly, evolving, and impacting numerous areas of life, with prevalence in sectors such as wearable technologies, smart sensors, and home devices. IoT networks are vulnerable to intrusion because dif- ferent devices use protocols and have limited computing power. These technologies are susceptible to hacking due to their rapid expansion, which poses significant security risks to user privacy and data protection, particularly given the increasing complex- ity of cyberattacks. We suggest an Intrusion Detection System (IDS) to address these vulnerabilities. Designed for IoT networks to detect and classify attacks. The system selects features using a genetic algorithm, and the GA fitness function was determined using a random forest model. The RF is used for intrusion detection processes. Utilize the Ton-IoT heterogeneous dataset. The recommended approach provides a robust defense against security risks associated with IoT. The testing results showed that a feature vector with 19 features achieved 99.59% multi-classification accuracy and 99.70% binary classification accuracy.