Ensemble Machine Learning Techniques for Attack Prediction in NIDS Environment

Abstract

The need for network intrusion detection systems (NIDS) to protect against different attacks growsas the scale of cyber attacks increases. The main areas of cyber attack research are its detection and prevention.Traditional ma- chine learning (ML) algorithms with low accuracy are used by the current NIDS, but it is not suitablefor newer anonymous cyber attacks. In this paper, an NIDS model with ensemble ML methods, which can detect andprevent different types of attacks compared with traditional ML methods, is proposed. Our specific sys- tem detectsknown attacks and blocks unknown attacks. The selected system uses four different machine learning methods,including data processing techniques for data preprocessing and data labeling. The entire NSL-KDD database isused to evaluate the performance of various ML classifiers based on different param- eters. The simulation analysisshows that the developed NIDS system is better than the existing single ML methods. The detection accuracy rate ofintrusion detection system (IDS) is increased by the model, which is essential for NIDS