Search

Aim to find effective methods for detecting and preventing obesity at early age. Material and methods. A dataset including the risk factors for child obesity was processed with artificial neural networks (ANN) and Statistica Neural Networks software. Clinical observations of 30 patients were used. The neural network was trained to predict the risk of obesity in children depending on the values of the selected parameters: standard deviation of body mass index from the norm, sex, age, obesity in parents, birth weight, duration of breastfeeding, deviation of body fat tissue content from the norm, and deviation of nutrition calories from the recommended values. Results. After training, the neural network MLP-8-7-1 was selected due to its high coefficients of determination 0.999999; 0.999407; 0.984930 for the training, test and control samples, respectively. This indicates the high performance of the trained ANN, the adequacy of which was checked graphically by constructing a histogram of residuals the difference between the entered and received by the network values of the risk of obesity development in children. Conclusion. The trained neural network can be used to predict the degree of risk of obesity in children and develop the necessary preventive measures in patients from risk groups.