A Whale Optimization Algorithm Feature Selection Model for IoT Detecting Intrusion in Environments

The Internet of Things (IoT) propagation has raised severe security concerns. Thus, the intrusion detection system (IDS) received enormous attention due to its critical in maintaining these environments’ security. Furthermore, various deep learning (DL) models were proposed to enhance the performance of the IDS in the literature. Hence, this paper proposed an IDS for IoT environments to increase the protection of the IoT environment. We applied Radial Basis Function Neural Network (RBFNN) as a multiclass classifier for the detection phase and a Whale Optimization Algorithm for the feature to improve the IDS performance. For the evaluation phase, we relied on the NF-ToN-IoT and NF-Bot-IoT datasets. Our model has scored significant results with 96.83% accuracy (ACC) and 89.74% Matthew’s correlation coefficient (MCC) on the NF-ToN-IoT, 98.43% ACC, and 57.71% MCC on the NF-Bot-IoT and 95.93% ACC and 82.68% MCC on the NF-ToN-IoT and NF-Bot-IoT dataset merged. Our model has shown outstanding results compared with other models.