Analysis of Intrusion Detection Using Machine Learning Techniques
G. Janani Alias Pandeeswari, S. Jeyanthi · 2022
The Internet of Things relates to many sensors and devices connected via the Internet. With IoT, data is easily accessible and can be exchanged with each other. Technological developments raise security problems. As these IoT devices are inter linked to sensors that are coupled to large cloud servers, smart city network traffic through IoT systems is increasing rapidly, creating new cyber security challenges. The increasing network traffic reflects in the task of identifying IoT attacks and detect malicious data at an early stage. The proliferation of IoT systems comes with the hazard of network attacks such as denial of service (DoS) and spoofing. Lot of experimenter have worked on IDS development using machine learning approaches to rectify the problems outlined above. With the accuracy of computer drills, the non-uniform data can be spotted automatically. Even unknown threats can be detected by machine learning systems due to their generalizability. Although several studies have addressed the application of machine learning (ML) solutions to detect attacks in recent years, little attention has been paid to detecting attacks on IoT networks. This paper contributes in this area by analysing varying machine learning methods to discover network attacks quick and perfectly. In this study, ADFA dataset is used for intrusion detection, intelligent anomaly detection based on the Random Forest machine learning algorithm, to address IoT cyber security and compare it with the literature study, and the new features have proven useful.