IoT-inspired smart home based urine infection prediction

Urine infection (UI) has been the most prevalent diseases among the global population. With large emphasize on smart healthcare in modern era, UI monitoring has been a major concern for medical industry. Motivated from these aspects, this paper proposes a novel Internet of Things-inspired framework in the form of a smart UI monitoring and prediction system for regular analysis of UI in home-centric environment. The overall framework has been structured in 5-layered model for determination of UI at early stages. These layers include urine perception layer, urine analysis layer, urine extraction layer, Urine prediction layer and visualization layer. These layers are designed to provide effective UI monitoring and analysis in real-time based on mathematical quantification of UI parameters in terms of infection degree and infection index measure for early prediction of UI. For prediction purposes, temporal-artificial neural network model is proposed. The visualization of the UI can be performed in time-sensitive manner over a handheld device for feasible analysis by the user. The validation of the proposed framework was done over real-time data of 12 patients acquired from nearby clinical laboratories. Results were compared with several state-of-the-art prediction approaches which show that the presented technique attain significant enhancement and high efficacy.