Suspicious Activity Detection and Classification in IoT Environment Using Machine Learning Approach
Veera Talukdar, Dharmesh Dhabliya, Bhupendra Kumar, Suryansh Bhaskar Talukdar, Shahanawaj Ahamad, Ankur Gupta
Abstract
Studies combining machine learning with the Internet of Things are now under process. The procedures for detecting fires are also described in depth. In order to be of any help, machine learning has to be able to identify and track the development of events that are equivalent to fires. The number of photos that must be analyzed by machine learning before a pattern may be found is decreased when data is captured and processed frame by frame. This information has been gathered for the express purpose of tracking a particular incident. Recent research has been given substantial treatment in this journal. Some of the constraints of the current state of the field have been discussed. The problem is identified, and suggestions on how to address it are provided. Using machine learning and image processing, we looked at one potentially suspicious IoT device. Many researchers have shown that the machine learning method can accurately classify and predict data. The authors of the present work have integrated machine learning with a compression method to further improve the security of the IoT ecosystem. Additionally, machine learning employs filtered datasets to improve the accuracy of attack classification. As a result of recent studies, it is possible to identify and classify suspicious events like fires, fake people, and other anomalies. Limiting the spread of misinformation relies on the ability to identify suspicious behavior.