Enhancing Network Forensic and Deep Learning Mechanism for Internet of Things Networks

Abstract

The integration of intelligence into everyday products has been possible due to the ongoing shrinking of hardware and a rise in power efficiency. The Internet of Things (IoT) area arose from the tendency to add computational capabilities to so-called non-intelligent daily items. IoT systems are attractive targets for cyber-attacks because they have many applications. Adversaries use a variety of Advanced Persistent Threat (APT) strategies and trace the source of cyber-attack events to safeguard IoT networks. The Particle Deep Framework (PDF), which is proposed in this study, is a novel Network Forensics (NF) that encompasses the digital investigative phases for spotting & tracing attack activity in IoT networks. The suggested framework contains three novel functionalities for dealing with encrypted networks, such as collecting network data flows & confirming their integrity, using a PSO algorithm, "Bot-IoT "& "UNSW NB15" datasets. The suggested PDF is related to several deep-learning methods. Experimental outcomes show that the proposed framework is very good at discovering & tracing cyber-attack occurrences when compared to existing approaches. The proposed design is implemented using neural network technology. The proposed design has 10% accuracy when compared with the existing structure. This paper is expected to offer a quick reference for researchers interested in understanding the use of network forensics and IOT.