Cloud based predictive data analysis framework for wearable device health alert system using semantic web services

Abstract

Rapid Innovation in Digital Technology achieved its frontier with fitness wearable technological devices. The ubiquitous tracking devices currently available in the market only monitor the amount of calories burnt by the user. They do not predict nor encourage users. This paper intends to provide prediction of calories burn based on users' physical activities, and encourage them to achieve more of their fitness goals, with the help of machine learning algorithms and ontology. The proposed framework has two different ontologies used for semantic synchronization. Fitness activities ontology deals with the predicted calories burn value and cloud Telephony ontology provides multi-channel alert services to the end user. FitBit Wearable fitness devices user data are analyzed from the cloud storage via cloud API, is proposed to interact with the user continuously with calories burn value for the improvement of their physical Activities like walking, jogging and step count. A custom model is constructed for predicting the calories burn value using Linear Regression Analysis through Machine Learning Algorithm. The proposed novel framework interacts with semantic web service registry through OWL API with the obtained predicted calories burn value from the prediction models. When compared to the existing system, the proposed framework produces enhanced insights on amount of calories burn to the user based on their activities through cloud telephony alerts like SMS, IVR, Mobile App and Email. The end user improves their activities from the obtained predicted value insights.