Determination of WHtR Limit for Predicting Hyperglycemia in Obese Persons by Using Artificial Neural Networks

The abdominal obesity is strongly associated with increased risk of obesity-related cardiometabolic disturbances. The proportion of waist circumference and body height, known as waist-toheight ratio (WHtR), has been shown as a good risk indicator related with abdominal obesity. This paper presents a solution based on artificial neural networks (ANN) for determining WHtR limit for predicting hyperglycemia in obese persons. ANN inputs are body mass index (BMI) and glycemia (GLY), and output is weist-to-height ratio (WHtR). ANN training and testing are done by dataset that includes 1281 persons.