An Effective Feature Selection Model Using Hybrid Metaheuristic Algorithms for IoT Intrusion Detection.

Area of Science:

• Cybersecurity
• Artificial Intelligence
• Data Science

Background:

• The proliferation of Internet of Things (IoT) applications generates vast amounts of data, increasing security vulnerabilities.
• Current security measures and intrusion detection systems (IDS) struggle to cope with sophisticated cyberattacks in IoT environments.
• Effective feature selection (FS) is crucial for optimizing machine learning algorithms in IDS, improving accuracy and efficiency.

Purpose of the Study:

• To develop a novel and effective feature selection (FS) method to enhance the performance of artificial intelligence (AI)-based intrusion detection systems (IDS) for Internet of Things (IoT) security.
• To improve the classification accuracy and convergence speed of AI algorithms used in IoT security by optimizing the feature selection process.
• To introduce a hybrid optimization approach, GTO-BSA, by integrating the Gorilla Troops Optimizer (GTO) with the Bird Swarm Algorithm (BSA) for superior feature selection.

Main Methods:

• A new feature selection (FS) method, GTO-BSA, was developed by enhancing the Gorilla Troops Optimizer (GTO) using the Bird Swarm Algorithm (BSA).
• The GTO-BSA method was designed to improve the exploitation capabilities of GTO, leading to better identification of optimal solutions and enhanced convergence.
• The performance of the GTO-BSA method was rigorously evaluated on four diverse IoT-IDS datasets: NSL-KDD, CICIDS-2017, UNSW-NB15, and BoT-IoT.

Main Results:

• The proposed GTO-BSA feature selection method demonstrated a superior convergence rate compared to the original GTO and BSA algorithms.
• Experiments showed that GTO-BSA achieved higher-quality solutions in feature selection, leading to improved performance metrics for IoT intrusion detection.
• Comparative analysis against state-of-the-art techniques confirmed the effectiveness and efficiency of the GTO-BSA approach on multiple benchmark datasets.

Conclusions:

• The GTO-BSA feature selection method offers a significant advancement in AI-driven IoT security, particularly for intrusion detection systems.
• The hybrid GTO-BSA approach effectively addresses the challenges of feature selection, leading to enhanced accuracy and faster convergence in cyberattack detection.
• This research provides a promising solution for improving the robustness and performance of security systems in the rapidly expanding landscape of Internet of Things applications.