Cyber Attacks Detection using Machine Learning

The Internet of Things, sometimes known as IoT, is a network that makes it possible for inanimate objects to communicate with one another virtually entirely independent of the intervention of humans. The number of things that can be connected by this network is in the hundreds of millions. The Internet of Things (IoT) is one of the computing disciplines that is increasing at one of the fastest rates; nevertheless, the harsh reality is that the internet is a very hostile environment, which makes the IoT vulnerable to a wide variety of different kinds of attacks. One of the realistic defences that must be created to protect IoT networks and address this problem is network anomaly detection. While preventing all attacks is challenging in the long run, successful defence in the present depends on the early discovery of an attack Because IoT devices have limited storage and processing, traditional high-end security solutions cannot secure them. Additionally, devices connected to the Internet of Things may now maintain a connection for longer periods of time without human input. Therefore, intelligent network-based security solutions, such as those based on machine learning, are necessary. The application of machine learning strategies to problems relating to attack detection has been the primary focus of a significant amount of research that has been published over the course of the past several years. This study has been carried out over a period of several years. Nevertheless, this is contingent upon the identifying of the assaults, especially those targeted at Internet of Things (IoT) networks, has received very little attention. By examining different machine learning algorithms that are capable of promptly and effectively identifying assaults on Internet of Things networks, this study seeks to further knowledge. Using a brand-new dataset named Bot-IoT, the effectiveness of several detection techniques is evaluated. Seven distinct machine learning algorithms were tried throughout the implementation process, and most of them were successful in achieving high performance. The Bot-IoT dataset's most recent properties were taken out and used during installation. The results of the new features were superior when they were compared to studies from previously published research.