A rapid discrimination of diabetic patients from volunteers using urinary volatile and an electronic nose

The electronic nose is a well-known technology able to classify differences between patients and healthy humans. This study investigates an Alpha M.O.S electronic nose for discrimination between diabetic patients and healthy volunteers using urinary aromas. In total 91 urine samples are used in this study, of which 43 belong to patients with type II diabetes and 48 belong to volunteers. Feature extraction is performed using Independent component analysis (ICA), with a random forest classifier, enabling 100% accurate classification of samples held out from training (AUC = 1)