Data dissemination approach using machine learning techniques for WBANs

Abstract

SummaryNowadays, the technological advancements of low power electronic and sensing devices, wearable systems, communication technologies, and cloud computing have encouraged the provision of ubiquitous health monitoring, medical diagnosis, and treatment consultations. The data collected from the patients are transmitted to mHealth server for real‐time and self‐reliant activity detection, behavior analysis, ambient assisted living, elderly care, activity of daily living, rehabilitations, entertainments, and surveillance in smart home environments. This helps in building sensor analyst systems that analyze the data on mHealth server for disseminating it to the respective end users. This is of utmost importance as it can provide real time feedback to patients, family members, caregivers, and so forth about the behavioral changes of elderly people and people with special needs. In this article, we have proposed knowledge based data dissemination system using machine learning techniques that analyses the data collected on mHealth server in three steps. First, the data are analyzed using support vector machine, k‐nearest neighbor, neural network, and logistic regression. Second, it finds hidden patterns from the data using hidden Markov model (HMM). Third, it defines fuzzy rules to disseminate the data to the end users. The system has been compared against conventional approaches. The uniqueness of the system lies in the fact that it uses temporal nature of data and provides with real‐time feedback as well as predicts outcomes and detects hidden patterns. The results have shown that the techniques NN, HMM, and fuzzy logic when used in conjunction with each other for disease prediction, hidden pattern detection, and data dissemination gives an accuracy of 98%. Thus, increasing effectiveness of sensor analyst system.