HSDL-based intelligent threat detection framework for IoT network

Abstract

Many Internet of Things (IoT) devices are susceptible to cyber-attacks. Attackers can exploit these flaws using the internet and remote access. An efficient Intelligent threat detection framework is proposed for IoT networks. This paper considers four key layout ideas while building a deep learning-based intelligent threat detection system at the edge of the IoT. Based on these concepts, the Hybrid Stacked Deep Learning (HSDL) model is presented. Raw IoT traffic data is pre-processed with spark. Deep Vectorized Convolution Neural Network (VCNN) and Stacked Long Short Term Memory Network build the classification model (SLSTM). VCNN is used for extracting meaningful features of network traffic data, and SLSTM is used for classification and prevents the DL model from overfitting. Three benchmark datasets (NBaIoT-balanced, UNSW-NB15 & UNSW_BOT_IoT- imbalanced) are used to test the proposed hybrid technique. The results are compared with state-of-the-art models.