Diabetes a common ailment affecting the vast population of people requires continues monitoring of blood glucose levels so as to control this disorder. Presently, the common technique used to monitor these levels is through an invasive process of drawing blood. Although this technique achieves high accuracy, it encompasses all disadvantages associated with an invasive method. This inconvenience is felt more accurately in patients who frequently examine these levels through the day. Hence, there is a need for a noninvasive technique for predicting the glucose levels. This paper aims at analyzing the breath as a noninvasive technique to predict diabetes. © Springer Nature Singapore Pte Ltd. 2018.