IoV Cyber Defense: Advancing DDoS Attack Detection with Gini Index in Tree Models
Muhammad Dilshad, Bushra Almas, Noshina Tariq, Husin Jazri, Ghadah Naif Alwakid, Jan Sher Khan, Pranav Kumar, Rahul Kumar · 2024
In today's ever-changing world of cybersecurity, where smart devices and the Internet of Vehicles (IoV) are everywhere, it is crucial to have a strong defense against attacks. This paper introduces an intelligent way to detect the Distributed Denial of Service (DDoS) attacks, a severe threat that can disrupt access to essential resources. Traditional methods to handle such attacks include intrusion detection systems (IDS). However, we need new and creative solutions to increasingly complex data. Our proposed strategy is an Artificial Intelligence (AI)-based technique that uses a Decision Tree (Dt) algorithm with feature selection using the Gini index. The proposed DDoS detection system has depicted an impressive accuracy of 99.4% while tested with the CIC-DDoS2019 dataset, surpassing baseline and previous benchmarks like Random Forest and XGBoost. Our improved Gini index method smartly selects 39 essential security features from a pool of 88, cutting down on unnecessary data and preventing over-fitting. This approach efficiently demonstrates its practical use in real-world network security situations, especially IoVs.