Adaptive Inertia Weight with Transient Search Optimization Based Feature Selection for Intrusion Detection in Internet of Things

The rapid growth of Internet of Things (IoT) devices across various sectors is continuously fuelled by increasing demand for smart technologies and enhanced interconnectivity.However, security is the main challenge because various devices utilizing different protocols and have limited computational capability cause vulnerabilities in IoT networks.The Adaptive Inertia Weight (AIW) with Transient Search Optimization (TSO) based feature selection approach is proposed to choose appropriate features for classification.Then, the Gated Recurrent Unit (GRU) with Derivative Sigmoid Weighted Linear Units (dSiLU) method is used to classify the types of intrusions with high detection rates.This process minimizes the vulnerabilities and enhances security in the IoT networks.The performance of AIW with TSO and GRU methods is evaluated with standard datasets of ToN-IoT, UNSW-NB15, N-BaIoT, Pre-IoT, CIDDS and NSL-KDD datasets.The AIW with TSO and GRU method reached 99.32% accuracy on the ToN-IoT dataset and 98.79% accuracy on the UNSW-NB15 dataset.The proposed approach effectively detects the intrusion and improves the security in IoT networks when compared with Whale combined Long Short-Term Memory (WILS).