Performance evaluation of machine learning algorithms with fuzzy logic for intrusion detection in VANET network
نویسندگان
1 Department of Computer Engineering, College of Engineering Kashibai Navale, Vadgaon (Bk.) Pune, Savitribai Phule Pune University, Pune, India.
2 Department of Computer Engineering, College of Engineering Kashibai Navale, Vadgaon (Bk.) Pune, Savitribai Phule Pune University, Pune, India.
3 Department of Computer Engineering, College of Engineering Kashibai Navale, Vadgaon (Bk.) Pune, Savitribai Phule Pune University, Pune, India.
4 Department of Information Technology, College of Engineering PVG’S, Technology and Management, Pune, Savitribai Phule Pune University, Pune, India.
doi
10.22105/jfea.2025.505777.1790چکیده
In recent years, Vehicular AD-HOC Networks (VANETs) have become a cornerstone in the advancement of Intelligent Transportation Systems (ITS), facilitating real-time communication between vehicles and infrastructure. While VANETs offer promising advancements in transportation, their inherent vulnerability to cyber threats makes it crucial to implement effective Intrusion Detection Systems (IDS). This paper proposes a novel approach to enhancing IDS performance in VANETs through fuzzy-based feature selection optimization. By integrating fuzzy logic into the feature selection process, the proposed method identifies and prioritizes the most relevant features for intrusion detection and significantly reduces computational overhead. The current proposed work utilizes Light Gradient Boosting Machine (LGBM) with Fuzzy Logic, achieving an impressive accuracy of -0.987, a precision of 99%, recall of 98%, and an F1-score of 0.99, demonstrating high accuracy and efficient processing time. A comprehensive evaluation using real-world VANET data demonstrates the superiority of the proposed system over traditional feature selection methods. The results indicate a significant enhancement in IDS performance, contributing to the broader field of vehicular network security.
