A Framework For Comparing Different Machine Learning Algorithm Models For Intrusion Detection In loT Environment

Abstract

loT-based systems can be seen in different areas like healthcare, transportation, farming, the power infrastructure, and manufacturing. Even though the Internet of Things can make people's lives easier, its exponential growth makes it a popular target for cyber-criminals and is subject to significant threats. One of the most devastating attacks is the denial of service (DoS), which prevents legitimate users from accessing services they have paid for. Therefore, There is an urgent requirement of loT-specific intrusion detection systems to tackle all these cyber-attacks. Numerous lightweight protocols are there to secure the communication between the loT devices. Here we used the IDS data set of the most critical loT communication protocol known as Message Queuing Telemetry Transport (MQTT) The information will be used as the foundation for developing creative intrusion detection method in loT networks. This work focused on developing a framework to compare several machine learning algorithms, and display the performance result of each one. The result demonstrated the most accurate model and the importance of using the machine learning-based IDS.