Intelligent remote health monitoring using evident-based DSS for automated assistance

The shift from common diagnosis practices to continuous monitoring based on body sensors has transformed healthcare from hospital-centric to patient-centric. Continuous monitoring generates huge and continuous amount of data revealing changing insights. Existing approaches to analyze streams of data in order to produce validated decisions relied mostly on static learning and analytics techniques. In this paper, we propose an incremental learning and adaptive analytics scheme relying on evident data and rule-based Decision Support System (DSS). The later continuously enriches its knowledge base with incremental learning information impacting the decision and proposing up-to-date recommendations. Some intelligent features augmented the monitoring scheme with data pre-processing and cleansing support, which helped empowering data analytics efficiency. Generated assistances are viewable to users on their mobile devices and to physician via a portal. We evaluate our incremental learning and analytics scheme using seven well-known learning techniques. The set of experimental scenarios of continuous heart rate and ECG monitoring demonstrated that the incremental learning combined with rule-based DSS afforded high classification accuracy, evidenced decision, and validated assistance.