Paper #2

Enhancing Intrusion Detection System Performance Using a Modified Grey Wolf Optimizer

Abdullah Al Mosuli, Mosleh Abualhaj, Ahmad Abu-Shareha, Mohamed Yousif and Mohammad Daoud

Abstract: Cybersecurity is one of the main worries of organizations, businesses, and even individuals. The problems facing cybersecurity are increasing on daily basis as a result of the increased reliance on electronic services and technologies and the associated increase in the number of cyberattacks. The prevention of cyberattacks has become a serious challenge due to the vast increase in cybersecurity threats. Intrusion Detection System (IDS) acts as one of the first line of defence against cyberattacks, protecting computer networks and users’ data. However, the efficiency and effectiveness of IDS can be challenged by the enormous data monitored by the IDS, and the irrelevant features in the data. This study presents a Machine Learning (ML) model for intrusion detection and aims to enhance the model by employing the proposed Modified-Grey Wolf Optimizer (GWO) for feature selection. A new mutation function and an effective initialization method are introduced to the GWO, enhancing its exploration of the solution space and reducing convergence time. The proposed modified-GWO is then applied to the NSL-KDD dataset for feature selection, identifying the most relevant features for intrusion detection. The ML model will be tested using various ML classifiers. These classifiers are XGBoost, RF, HGB, and MNB. The proposed model achieved remarkable results with the XGBoost classifier reaching an accuracy of 99.52%, a precision of 99.47%, and a recall of 99.46%.

Keywords: Feature Selection; Grey Wolf Optimizer; Machine Learning; NSL-KDD Dataset.