1. Securing IIoT operations with recurrent federated network-based enhanced local search grasshopper.
- Author
-
Alassafi, Madini O.
- Subjects
- *
FEDERATED learning , *OPTIMIZATION algorithms , *INDUSTRIAL productivity , *INDUSTRIAL efficiency , *INTERNET of things - Abstract
The integration of the Industrial Internet of Things (IIoT) brings about a significant improvement in the efficiency and productivity of industrial processes. The speed and accuracy of various tasks have been greatly enhanced with the utilization of IIoT which allows smoother operations and cost-effective solutions, to protect uninterrupted intrusions and avoid potential threats. This article proposes a novel recurrent federated network-based enhanced local search grasshopper (RFN-ELG) algorithm. Datasets like UNSW-NB15 and MQTT-IoT-IDS2020 datasets are employed to determine the performances of IIoT via two diverse phases, namely the data preprocessing phase as well as attack detection phase. The recurrent neural networks (RNNs) are integrated with federated learning (FL) to overcome the gradient issues during the training process. In addition to this, the hyperparameters of RNN-FL are tuned via a grasshopper optimization algorithm with a local search strategy. Accuracy, precision, recall, F1-score, and false alarm rate (FAR) are the performance metrics taken for attack detection, where accuracy attains the value of 98.1%, precision attains the value of 97.4%, recall attains the value of 96.1%, F1-score attains the value of 97.2%, and FAR attains the value of 96.1%, respectively. Therefore, the proposed RFN-ELG algorithm attains high performance in detecting the attacks in IIoT. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF