Enhanced IoT Based IDS Driven by Binary Snake Optimizer for Feature Selection.

Connected devices have extended the borders of the traditional Internet into the new Internet of Things (IoT). IoT holds a significant role in several fields such as industry, transportation, smart homes, cities, and others. However, protecting IoT environments and preventing intrusions is one of the critical problems in IoT. An intrusion detection system (IDS) aims to identify malicious patterns and threats that traditional security countermeasures cannot detect. This paper presents an effective feature selection (FS) approach driven by a binary variant of the newly proposed Snake Optimizer (SO) for enhancing intrusion detection systems. Two variants of FS are developed, and the best one that is based on a V-shaped transfer function is compared with several optimization algorithms to confirm its efficiency in boosting IDSs. In addition, five datasets that represent real IoT traffic are employed for evaluation purposes. The experimental results show that SO based on V-shaped transfer is superior to the S-shaped transfer function and outperforms other optimizers in particular based on the obtained average accuracy and convergence rates. Hence, it can conclude that the proposed approach can be efficiently employed in IoT intrusion detection systems.