Machine Learning based Attacks Detection and Countermeasures in IoT

Abstract

While the IoT offers important benefits and opportunities for users, the technology raises various security issues and threats. These threats may include spreading IoT botnets through IoT devices which are the common and most malicious security threat in the world of internet. Protecting the IoT devices against these threats and attacks requires efficient detection. While we need to take into consideration IoT devices memory capacity limitation and low power processors. In this paper, we will focus in proposing low power consumption Machine Learning (ML) techniques for detecting IoT botnet attacks using Random forest as ML-based detection method and describing IoT common attacks with its countermeasures. The experimental result of our proposed solution shows higher accuracy. From the results, we conclude that IoT botnet detection is possible; achieving a higher accuracy rate as an experimental result indicates an accuracy rate of over 99.99% where the true positive rate is 1.000 and the false-negative rate is 0.000.