Using Machine Learning for Anticipating a Diabetes Crisis Through a Sensors-based Internet of Bio-nano Things Network

Abstract

It is shown that Machine Learning can be exploited in medical urgencies such as a diabetic crisis, that is manifested in showing high values in both glucose and blood pressure. Often patients cannot estimate a possible crisis so that, in most cases most of them are quite sensitive to strokes and cardiac arrest unexpectedly. In this paper the algorithm of Tom Mitchell is employed as a kind of software that manages the updated inputs in order to anticipate a possible crisis in terms of probabilities. Thus, while sensors are enough accurate to measure a variable, the algorithm is able to make predictions about the worse scenarios of diabetes crisis. When this information is monitored inside an Internet of Bio-nano Things network, patients might be assisted in the shortest times, by avoiding irreversible complications in their health. Therefore, a health services operator acquires capabilities to minimize risks and make fast and precise decisions with minimal errors either from clinicians and instruments.