An Intrusion Detection and Classification System for IoT Traffic with Improved Data Engineering

Nowadays, the Internet of Things (IoT) devices and applications have rapidly expanded worldwide due to their benefits in improving business environment, industrial people’s daily lives. However, IoT are not immune malicious network traffic, which causes potential negative consequences sabotages operating devices. Therefore, developing a method for screening traffic is necessary detect classify activity mitigate its impacts. This research proposes predictive machine learning model an system. Specifically, our distinguishes between normal anomaly activity. Furthermore, it classifies into five categories: normal, Mirai attack, denial service (DoS) Scan man-in-the-middle (MITM) attack. Five supervised models were implemented characterize performance detecting classifying activities systems. includes following models: shallow neural networks (SNN), decision trees (DT), bagging (BT), k-nearest neighbor (kNN), support vector (SVM). The evaluated on new broad dataset attacks, IoTID20 dataset. Besides, deep feature engineering process was used improve models’ accuracy. Our experimental evaluation exhibited accuracy 100% recorded detection using all 99.4–99.9% classification process.