IoT and Machine Learning based Self Care System for Diabetes Monitoring and Prediction

Diabetes is a chronic disease caused by the assimilation of blood sugar, mainly because of reduced production or no production of insulin within the body (type 1 diabetes), or because cells are irresponsive to the produced insulin (type 2 diabetes). In recent years, a multitude of people turned out to be diabetic and is increasing drastically. Moreover, a report by World Health Organization describes 346 million people are affected by diabetes around the world. Furthermore, the lack of a self-care system for monitoring and detecting signs at an early stage in the patient’s data causes pre-diabetes or diabetes condition which remains unrevealed in more than one-third of the population and later diagnosed with diabetes. The combination of machine learning techniques and the Internet of Things can provide an effective solution to predict diabetes well before. Therefore, this paper presents an Internet of Things (IoT) and Machine Learning-based non-invasive self-care system which monitors blood sugar and various vital parameters to predict diabetes well before. The non-invasive way of measuring blood sugar through a developed IoT sensor is much more comfortable compared to the invasive method. In the proposed system deployment of the SVM-based machine learning model on the cloud and its integration with the android application enables doctors and patients to monitor the vital parameters and associated risk easily. In addition to this, monitored parameters are sent to the doctor through email for further analysis, and suggestions in diet and lifestyle based on the monitored parameters are conveyed to the patient through an android application to prevent or reduce the risk of diabetes. Thus, the proposed self-care system can overcome challenges of the traditional way of monitoring diabetes and helps patient and doctor in monitoring, recording, and analyzing data for the prognosis of diabetes.