SmartHealth: Machine Learning based Smart Diabetes Prediction in an IoT System

The development of machine learning and Internet-Of-Things has helped to break down barriers in health and medical facilities. In this paper, we propose SmartHealth, a smart diabetes prediction framework in an Internet-Of-Things (IoT) system based on machine learning. The factors (attributes) affecting diabetes are gathered from individuals using a collection of IoT devices and recorded in our designed SmartHealth mobile application. We design the machine learning models using different algorithms, such as k-Nearest Neighbour, Logistic Regression, Random Forest, Decision Tress, and Multi Layer Perceptron and determine the best performing model among them. This model has been pre-trained and is used to predict the result of the individuals based on the collected data (attributes) that is uploaded in the edge server. Different results and plots are obtained that convey the performance of our SmartHealth framework.