An IoT Network Identifier Using Machine Learning-Assisted Feature Selection and SMOTE

The Internet of Things (IoT) is a network of interconnected physical devices and things that gather, process, and distribute data using sensors, actuators, computers, and network nodes. With the proliferation of the Internet of Things (IoT) devices in recent years, businesses have opened themselves up to a more significant number of potential security flaws and attacks.Therefore, institutions must identify how many and what kinds of devices are connected to their network and whether or not they constitute a security concern. Feature information is first retrieved by intercepting device data and then exploited to offer device classification. This method has gained favour in recent years for identifying devices via networks using supervised learning. Existing efforts predominantly suffer from a lack of scalability due to their reliance on pattern matching to identify devices; any newly introduced device kinds are automatically labelled as anomalous. Each year, millions of new Internet of Things devices are created in the real world. In this research, a supervised learning approach for automatically classifying devices that have not been manually tagged is offered. The device proposed in this paper incorporates a complete classification of many ensemble balance classification strategies, which may be used to categorize devices as either authorized or unauthorized.