Personalized remote health monitoring based on activities and Vital parameters using soft computing techniques

Abstract

Wireless technology, body area networks, artificial intelligence newlineprovides the freedom for the patients to be continuously monitored at any newlineplace at any time. Health parameters are highly uncertain, non-linear, and newlinedynamic in nature. This requires an adaptive learning process to adjust the newlinemodel behaviour as per the streaming data. Soft computing approaches are newlinestrongly recommended for handling these dynamic and complex real-world newlineapplications. Existing soft computing approaches failed in vital parameter newline(or) post-surgery conditions etc. newlineTo address these issues, this thesis proposes a personalized healthcare newlinescheme using soft computing techniques for detecting the abnormality. The newlinescheme involves activity recognition, personalization of vital parameter newlinevalues based on activities and health status, abnormality detection using the newlinepersonalized values and all these techniques are integrated with dynamic newlineservice scheduling in cloud based remote health monitoring. newlineThe research proposes Optimized ANFIS using Frequent Pattern newlineMining (OAFPM) for activity recognition, Density based K-Means Clustering newline(DbK-MeansC) for severity range fixation, a novel scheme based on Naïve newlineBayes (NB), Bayesian Belief (BB) and Genetic Algorithm (GA) - NB3GA for newlinepersonalization and abnormality detection. The integrated Dynamic Priority newlineScheduler (DPS) in cloud reduce the latency between the patient and doctor. newline