Malware Detection in IoT Network using DBSCAN and Ensemble method

Abstract

The rapid growth of the Internet of Things introduced several daunting security challenges, major ones being anomaly detection that may compromise the integrity of IoT networks. This paper introduces an anomaly detection system specifically engineered to enhance IoT security through Machine Learning techniques. The system utilizes the IoT-23 dataset for the purpose of finding anomalies in network traffic while employing rigorous data preprocessing to optimize traffic analysis. It has two phases. In the first phase, IoT network data undergoes an in-depth preprocessing phase, and then DBSCAN and Random Forest are applied to compare their results. It contains a traffic capture unit that captures data from the IoT sensors and sends them to the computer unit for capturing real-time anomalies. The experiments show that applying ML techniques to IoT anomaly detection leads to an almost highly efficient and accurate solution that makes it a suitable approach to further improve security in the confined resource environment of IoT