A New Model for Enhancing IoT Security Through Hybrid Optimization of Intrusion Detection
Manel Rafrafi, Cherif Ghazel, Leı̈la Azouz Saı̈dane · 2024
The Internet of Things (IoT) allows multiple smart devices to communicate with each other with minimal human intervention. IoT is currently an increasingly popular area of research due to its rapid growth and use in many areas including health, transport, smart cities, smart homes and others. However, IOT face several security threats and challenges. The implementation of traditional security measures. To address these issues, it's necessary to ensure the security, confidentiality, integrity and availability of data in the IoT. The use of Intrusion Detection Systems (IDS) is one of the fundamental solutions available to tackle these challenges. In this study, our objective is to enhance the IoT security by proposing and implementing a new optimized intrusion detection system capable of improving detection accuracy and reducing the rate of false alerts. We define a hybrid technique that combines optimization with classification algorithms in machine learning as well as in deep learning. The main purpose of this hybrid approach is to boost accuracy, precision and performance of the IoT intrusion detection systems while reducing false alerts. To verify security and assess the validity of our proposal, we conduct experiments to validate its accuracy, precision, robustness and correctness.