Artificial neural network applied in the analysis of health-related quality of life of adolescents

Objective: to construct a model that explains the health-related quality of life among school adolescents from the KIDSCREEN-27 instrument through the creation of an artificial neural network. Method: cross-sectional and analytical study with 635 adolescents using KIDSCREEN-27. An artificial neural network with four layers was developed to evaluate the variable health-related quality of life by means of the mean responses. For the first three layers of neurons, logistic function was used as transfer function and linear function was used for activation. Results: the neural network reached accuracy of 98.96% and when compared the dimensions of kidscreen-27 with sex and practice of physical activities all presented significant statistical association, except the dimensions social support and peer group and school environment. Conclusion: the results may have important consequences for the identification of adolescents at risk and the direction of public health policies.