Hybrid dual-channel convolution neural network (DCCNN) with spider monkey optimization (SMO) for cyber security threats detection in internet of things

Abstract

Applications for the Internet of Things (IoT) have been employed in a broad range of industries, including smart homes, healthcare, smart energy, and Industrial 4.0. IoT offers many advantages, such as ease and efficiency, but it also adds a lot of new risks. The problem is often made worse by the potential number of linked IoT devices and the ad hoc nature of such networks. IoT management now faces substantial difficulties related to safety and confidentiality. Current research has shown that deep learning algorithms are quite effective and have numerous benefits over previous approaches for doing security assessments of IoT devices. An integrated deep learning approach is created in this work to identify files containing malware and pirated software through the Internet of Things network The Hybrid Dual-Channel Convolution Neural Network (DCCNN) with Spider Monkey Optimization (SMO) is called DCCNN-SMO advocated leveraging stolen reference code to detect software piracy. For the purpose of examining software piracy, the dataset was gathered via Google Code Jam (GCJ). To research the suggested strategy, collected 100 input data for internet users from GCJ. In addition, the DCCNN-SMO is utilized to visualize colored images to identify harmful intrusions in IoT networks. From the Leopard Mobile database, malware samples were taken for testing. The actual results demonstrate that the recommended method for calculating the cyber security risks in IoT performs categorization more over current approaches. The proposed DCCNN-SMO + SVM approach provides better resultant value of 98.55% whereas the other existing approaches like GIST + SVM is 86.1%, CLGM + SVM is 92.06% and DNN + SVM is 97.46%, LBP + SVM is 78.05%.