A Modified Grey Wolf Optimization Algorithm for an Intrusion Detection System

Cyber-attacks and unauthorized application usage have increased due to the extensive use of Internet services applications over computer networks, posing a threat service’s availability consumers’ privacy. A network Intrusion Detection System (IDS) aims detect aberrant traffic behavior that firewalls cannot detect. In IDSs, dimension reduction using feature selection strategy has been shown be more efficient. By reducing data eliminating irrelevant noisy data, several bio-inspired algorithms employed improve performance an IDS. This paper discusses modified algorithm, which is Grey Wolf Optimization algorithm (GWO), enhances efficacy IDS in detecting both normal anomalous network. The main improvements cover smart initialization phase combines filter wrapper approaches ensure informative features will included early iterations. addition, we adopted high-speed classification method, Extreme Learning Machine (ELM), used GWO tune ELM’s parameters. proposed technique was tested against various meta-heuristic UNSWNB-15 dataset. Because generic attack most common type dataset, primary goal this attacks traffic. model outperformed other methods minimizing crossover error rate false positive less than 30%. Furthermore, it obtained best results with 81%, 78%, 84% for accuracy, F1-score, G-mean measures, respectively.