Processing model for fog computing applied to internet of medical things (IoMT)

Internet of Things (IoT) platforms are software systems that connect and manage IoT devices, applications, and data, act as intermediaries between devices and applications, providing a common framework for them to communicate with each other. Furthermore, in fog computing architecture, the fog layer refers to the layer of computing infrastructure that is located between the edge devices and the cloud data center. Therefore, fog layer is comprised of inter-mediate nodes, which are computing resources that are located closer to the edge devices than the cloud data center, allows to improve the efficiency and reduce the latency of data processing by bringing computing and storage closer to the devices that generate and use data. This study presents a new model of real-time data processing in the fog layer to take advantage of own resources instead of sending data to a centralized cloud data center for processing. The method used was supervised machine learning, which consists of creating a model that makes predictions using evidence supported by uncertainty applied in e-health, using intermediate nodes within a platform used for the detection of heartbeat anomalies in an Internet of Things (IoT) network. The evaluation used was a set of statistical validation metrics. The results may identify important criteria for choosing appropriate machine learning techniques, which may include the statistical and inherent effectiveness of ML methods or their adaptability to intermediate nodes within an IoT platform. Translated with DeepL.com (free version) the less resource-intensive methods like Simple Linear Regression, Logistic Regression, and K Nearest Neighbors are suitable for intermediate nodes, as they demand minimal processing and storage. Consequently, adopting a cognitive network approach for intermediate nodes in IoT platforms reduces the cloud computing processing costs and shifts the burden to the fog layer.