A Hybrid Approach for the Detection and Classification of MQTT-based IoT-Malware

In recent, Internet of Things (IoT) networks are more common because they can keep track of constantly changing environmental or network field conditions. It incorporates a significant level of internet access, smart device use, and data exchange. Since there are more connected devices in the IoT network, it is crucial to protect the IoT network’s security. Message Queuing Telemetry Transport (MQTT) is one of the most popular IoT communication protocols. In this study, a hybrid approach is applied to integrates two machine learning algorithms, namely multi-layer perceptron (MLPC) and K-Nearest Neighbour (KNN) for the detection of IoT malware based on MQTT protocol communication. It was simple to recognize or anticipate the attack to address the security risk of the MQTT protocol based on these ML methods. The results generated by the proposed model is 91% and 93% respectively.