| Popis: |
Cybersecurity continues to be a significant problem for all industries involved in digital activities, and it is specified as the cyclical surge in security occurrences. Considering more Internet of Things (IoT) devices are being employed in the medical field, homes, transportation, offices, and other locations, malicious attacks are arising more regularly. While IoT provides numerous advantages to users or service providers, security and privacy remain a significant problem. An Intrusion Detection System (IDS) can be employed to mitigate cyber threats in such an interconnected network. However, several existing performances for IDS in IoT need more extensiveness of the categories of attack the network was showing, higher-level feature dimensional, systems built on out-of-date databases, and a shortage of consideration of the imbalanced databases. Therefore, this study presents a Distributed Multiclass Cyberattack Detection using Golden Jackal Optimization with Deep Learning (DMCD-GJODL) technique for IoT networks. The main aim of the DMCD-GJODL method is to ensure security in the IoT environment by detecting cyberattacks using the DL model. In the DMCD-GJODL method, the min-max scalar is primarily used to scale the input data. The DMCD-GJODL method applies a Chaotic Crow Search Optimization Algorithm (CSSOA) based feature selection approach to select features. Moreover, the Bi-Directional Gated Recurrent Unit (BiGRU) approach can be exploited to detect and classify cyberattacks. Eventually, the GJO methodology can boost the BiGRU approach’s hyperparameter choice, enhancing the overall classification process. The performance evaluation of the DMCD-GJODL approach takes place on the BoT-IoT dataset. Extensive comparative results stated the betterment of the DMCD-GJODL approach in detecting cyberattacks with a maximum accuracy of 98.70%, precision of 98.92%, recall of 97.62%, and F-score of 98.25%. |
