Network Intrusion Detection System in Internet of Things Using Chaotic Elite Guidance Learning Strategy-based Lotus Effect Optimization Algorithm

Intrusion Detection Systems (IDS) are vital for Internet of Things (IoT) network security that is used to detect and prevent harmful actions like malicious attacks.The network intrusion data are combined into several typical examples because of its dynamic nature that results in a lack of instances to train the models and increase the false detection rate.To overcome these limitations, a Chaotic Elite Guidance learning strategy-based Lotus Effect Optimization algorithm (CEG-LEO) and Bi-directional Long Short-Term Memory (Bi-LSTM) with categorical crossentropy loss function is proposed for network intrusion detection in an IoT environment.The proposed CEG learning strategy with the LEO algorithm is used to select significant features based on the best fitness solution obtained by the pollination activity of the lotus flower.The proposed Bi-LSTM model is employed to detect and classify multiple intrusive attacks accurately by extracting temporal information from the selected attacks features which are in sequence order of network traffic events.The experimental results of proposed methods utilized in network intrusion detection in IoT achieved accuracy of 99.34 % and 99.37% for CICIDS-2018 and CICIDS-2019 datasets which is higher when compared to existing detection approaches like Rat Swarm Hunter Prey Optimization based Deep Maxout Network (RSHPO-DMN).