Published Paper Details:

Millions of devices with sensors and actuators connected via wired or wireless channels for data transmission make up the Internet of Things (IoT). By 2020, it is anticipated that over 25 billion devices will be connected, reflecting the IoT's tremendous growth over the last ten years. In the upcoming years, the amount of data released from these devices will multiply many-fold. In addition to producing more data overall, IoT devices also produce a lot of data in a variety of different modalities, with variable degrees of data quality determined by the speed of time and position dependency. In such a setting, machine learning algorithms can be crucial in assuring biotechnology-based security and authorisation as well as anomaly detection to enhance IoT systems. However, hackers frequently use learning algorithms to attack the flaws in IoT-based smart systems. In this study, we suggest employing machine learning to detect spam in order to secure IoT devices. Spam Detection in IoT utilising a Machine Learning framework is suggested to accomplish this goal. In this approach, a huge number of input feature sets are used to evaluate five machine learning models using a variety of criteria. Each model uses the enhanced input attributes to calculate a spam score. This rating shows how trustworthy Internet of Things (IoT) devices are based on several factors. The proposed technique is validated using the REFIT Smart Home dataset. In comparison to other current systems, the findings collected demonstrate the effectiveness of the proposed method.