Internet of Things: A survey on machine learning-based intrusion detection approaches

Abstract

In the world scenario, concerns with security and privacy regarding computer networks are always increasing. Computer security has become a necessity due to the proliferation of information technologies in everyday life. The increase in the number of Internet accesses and the emergence of new technologies, such as the Internet of Things (IoT paradigm, are accompanied by new and modern attempts to invade computer systems and networks. Companies are increasingly investing in studies to optimize the detection of these attacks. Institutions are selecting intelligent techniques to test and verify by comparing the best rates of accuracy. This research, therefore, focuses on rigorous state-of-the-art literature on Machine Learning Techniques applied in Internet-of-Things and Intrusion Detection for computer network security. The work aims, therefore, recent and in-depth research of relevant works that deal with several intelligent techniques and their applied intrusion detection architectures in computer networks with emphasis on the Internet of Things and machine learning. More than 95 works on the subject were surveyed, spanning across different themes related to security issues in loT environments. (C) 2019 Elsevier B.V. All rights reserved.