A Multi-Level Hybrid Classifier and Ensemble Feature Selection Based Intrusion Detection System

As the Internet and related technologies have advanced rapidly, so have the opportunities for infiltration and the number of hidden dangers to network security. Hence, in order to prevent attacks on networks, it is essential to improve intrusion detection systems' ability to identify and stop them. Consequently, researchers have come up with a number of ways to use network monitoring data to find network intrusions quickly and accurately. This has led to the use of network intrusion detection systems (NIDS). In this paper, a signature-based intrusion detection system (SIDS) using a multi-level hybrid system is introduced in order to detect and classify attack types. Experiments conducted on the UNSW-NB15 dataset indicate that the system under consideration has a detection capability of 99.87% for binary attacks and 93.68% for multiclass attacks. The results of the experiments show that the suggested multi-level hybrid approach can effectively find network intrusions and perform better than current intrusion detection systems (IDS).