Predicting Type 2 diabetes using an electronic nose-based artificial neural network analysis

摘要: Diabetes is a major health problem in both industrial and developing countries, its incidence rising. Although detection of diabetes improving, about half the patients with Type 2 are undiagnosed delay from disease onset to diagnosis may exceed 10 yr. Thus, earlier treatment hyperglycaemia related metabolic abnormalities vital importance. The objectives present study were examine urine samples diabetic healthy volunteers using electronic nose technology evaluate possible application data classification methods such as self-learning artificial neural networks (ANN) logistic regression (LR) comparison principal components analysis (PCA). Urine controls processed randomly simple 8-sensors individual patterns qualitatively classified "Approximation Classification Medical Data" (ACMD) network based on output neurons, binary LR PCA. Distinct classes found for subjects PCA, which had 96.0% successful percentage mean while qualitative ANN percentages 92.0% 88.0%, respectively. Therefore, ACMD suitable classifying medical clinical data.