IoTProtect: A Machine-Learning Based IoT Intrusion Detection System

Abstract

The rapid growth in IoT adoption in various daily-life applications, combined with the lack of proper patching and securing, has made IoT an easy target for malicious actors. As we notice the increase in the utilization of IoT devices in conducting security attacks around the world, research needs to catch up and protect IoT devices.In this paper, we present IoTProtect; a machine-learning based intrusion detection system utilizing the TON_IoT dataset in training and testing. Testing the proposed system showed 99.999% detection accuracy with 0.001% false-positive, and 0% false-negative with excellent timing performance.