Abstract

In the present healthcare climate, nearly 4 billion people lack access to medical care around the world. Diabetes is one of the major chronic disease many people are suffering from. Diabetes rates have doubled from 1980 to 2018, rising from 5% to 10% of the world population. The prevalence has specifically increased in low- and middle-income countries. The proposed system applies machine learning techniques to the collected dataset to improve diabetes progression technique, disease prediction and patient self-management. Machine learning techniques are widely used in this regard to develop analytic models. This system analyses the data given from the user to identify behavioral patterns and clinical conditions of the patient. This system analyses the collected data to identify the improvements in patient’s diabetic status, habits and anomaly in daily routines, change in sleeping and mobility, eating, drinking and digestive pattern. The user has to register providing personal data and present medication undergoing. The user can regularly log on to the system providing input data which are blood glucose level, blood pressure and present medications. The system processes the input using machine learning techniques and produces the output. The system determines whether the patient is normal or having diabetes on the basis of several inputs given by him or her. If the patient is having diabetes, it also determines which type of diabetes he or she is having. A graph will be plotted which shows the diabetic variation over a week, a suggestion file is given to the user which includes diet recommendations and further side effects patient may experience and measures to overcome them and a message is sent to the user’s phone number at the time he should take medications that he has mentioned earlier.