Detecting the Anomaly Attacks in Industrial IoT using Hybrid Feature Selection Based Modified Convolutional Neural Network

When it comes to digitally transforming classic sectors towards Industry 4.0, the Industrial Internet of Things (IIoT) is crucial. The IIoT improves the efficiency and productivity of businesses by connecting sensors, equipment used in industries to the Internet for the purposes of data collecting, data analysis, and automated control. Because of the intricate nature of the IIoT infrastructure, anomaly detection has emerged as a crucial method for guaranteeing the platform's viability. As IIoT is inherently suited to continuously growing networks, graph-level anomaly detection has proven to be an effective method for detecting and predicting abnormalities across a wide range of application domains, including transportation, energy, and the factory. This research work suggests an optimization strategy using a combination of the fuzzy technique for order of favorite by similarity to ideal solution (FTOPSIS). For network anomaly detection, the selected characteristics are fed into a modified Convolutional Neural Network (MConvNet). The approach is put through its paces using NBaIoT-balanced benign and malicious network traffic so that it can pick up on the differences between the two. When compared to state-of-the-art models, the findings show that the suggested model outperforms them thanks to the optimization approach's feature selection.