A hybrid Ant Lion Optimization algorithm based lightweight deep learning framework for cyber attack detection in IoT environment

Internet of Things (IoTs) are integral part of Web3, in which they are used for information collecting and sharing. However, the limited storage capacity of IoT decides made them vulnerable to many types of cyber attacks. In this context, we proposed a hybrid deep learning model for the detection of cyber attacks in the IoT environment. The proposed approach used features selection technique for the selection of efficient features and Ant Lion Optimization algorithm for tuning the hyper-parameters. This hybrid approach model train for five epochs and detects the attack traffic with an accuracy of 97%, which makes it efficient and lightweight for IoT applications. The proposed model is also outperformed the standard machine learning and deep learning models.