Aggregation in IoT for Prediction of Diabetics with Machine Learning Techniques

Abstract

Diabetes is a chronic illness and it may generate many dilemmas. Diabetes millitus patients in the earth will reach 650 million in 2050, which means that more number of adults will have diabetes in time ahead. There is no doubt that this startling number needs huge deliberation. Diabetic patients data are gathered and forwarded through Internet of Things (IOT). The lifespan of the network is the critical confront in the IoT. To elongate the lifespan of the IoT, data aggregation is a useful method to abate the number of transmissions among objects. Reduced number of data replication leads to elongate the network lifespan and to descent the energy depletion. The data collected from the diabetic patient are accumulated and the machine learning techniques are imposed to presage diabetics with a high degree of compassion and specificity. In this work, the K-Nearest Neighbor and Support Vector Machine are used to predict diabetes. The results showed that Support Vector Machine achieves the highest accuracy compared to K-Nearest Neighbor when all the attributes were used.