Deep neuro‐fuzzy approach for risk and severity prediction using recommendation systems in connected health care

Internet of Things (IoT) and Data science have revolutionized the entire technological landscape across the globe. Because of it, the health care ecosystems are adopting the cutting‐edge technologies to provide assistive and personalized care to the patients. But, this vision is incomplete without the adoption of data‐focused mechanisms (like machine learning, big data analytics) that can act as enablers to provide early detection and treatment of patients even without admission in the hospitals. Recently, there has been an increasing trend of providing assistive recommendation and timely alerts regarding the severity of the disease to the patients. Even, remote monitoring of the present day health situation of the patient is possible these days though the analysis of the data generated using IoT devices by doctors. Motivated from these facts, we design a health care recommendation system that provides a multilevel decision‐making related to the risk and severity of the patient diseases. The proposed systems use an all‐disease classification mechanism based on convolutional neural networks to segregate different diseases on the basis of the vital parameters of a patient. After classification, a fuzzy inference system is used to compute the risk levels for the patients. In the last step, based on the information provided by the risk analysis, the patients are provided with the potential recommendation about the severity staging of the associated diseases for timely and suitable treatment. The proposed work has been evaluated using different datasets related to the diseases and the outcomes seem to be promising