AN EFFICIENT MACHINE LEARNING AND DEEP LEARNING BASED DEEP PACKET SECURITY FRAMEWORK FOR DETECTION OF COMPUTING NETWORK FAULTS IN THE IOTS
Keywords:
Machine Learning, Deep Neural Network, CNN, Prediction Models, Routing Attacks Detection, Deep Learning, Internet Of Things, Threat Detection, Internet Of Things Networks, Wi-Fi security, wireless protocols, WEP, EncryptionAbstract
In Industry 4.0 deep learning plays vital role as Deep Learning for cybersecurity is applied in healthcare, the software industry and IoTs. Developing deep learning models is difficult, as both the world and the data are always changing. This article proposed an approach based on Deep Learning using an intrusion detection system. It also scrutinizes operations such as supervised operations and unsupervised ones. It introduces a new method for dealing with threats and achieves an accuracy rate of 71.73% in the l73 faulty packets. The article tries to find IoT systems, applications, data and services in the organization that might be vulnerable to cyber attacks due to their persistent connections. Software piracy and increased malware attacks are currently putting IoT security at risk. Once the information has been gathered, DL-IDS decides whether to send data to the fog layer. It demonstrates better results than other DL-IDS systems that were evaluated using the RT-IoT2022 dataset. The Intrusion Detection System with IPS-DL was found to be 71.73 percent accurate. When practicing intrusion prevention, the system’s accuracy was 70.63%, recall was at 96.30% and it achieved an F1-score of 92%. It can stop 85% of attacks and holds just 0.23% of the data lost from the sensors, with only 0.11 joules of power needed, as it keeps throughput high at 0.99% and its delivery ratio is ideal at 0.99%.
