Prioritization of IoT Devices Healthcare Data based on Attribute Scoring and Metadata Annotation

Current era is surrounded by data deriving from a plethora of platforms, sensors, and systems with Internet of Things (IoT) devices being among the top candidates that contribute towards this data explosion. As data evolves in variety, volume, and importance, additional focus should be provided on the most valuable data, since not all data are equally important and able to lead to valuable insights. Consequently, specifically designed techniques are implemented aiming to identify and take advantage of the subset of data that can provide meaningful and positive impact. Nevertheless, these techniques lack of generalization and applicability in multiple domains since they are aiming towards specifically designed scenarios. Towards this goal, this paper proposes a data prioritization technique through which incoming data of IoT devices can be filtered, offering the most critical slices of data to their requestors. The proposed approach is designed to prioritize the data to be sent from an IoT device of a citizen to a Healthcare Practitioner, where the first step involves the ingestion of these data, in order to be split, scored for their importance - based on already pre-trained classifiers, and annotated with metadata describing their nature. Following the attribute scoring and metadata annotation stages, these data are filtered and sliced based on specific requirements, being finally provided to the requesting Healthcare Practitioner. The latter approach has been experimented and evaluated with multiple datasets, efficiently managing and identifying IoT devices healthcare data of high and low priority.