Stochastic neural networks applied to dynamic glucose model for diabetic patients

In this paper, we have described the use of stochastic neural networks in the Bergman's model of Insulin-glucose interaction, this model is observable in the sense of control theory, the variables in the model cannot be measured on-line but these were estimated by the neural network. The variables behavior are presented for a typical input like a food ingest in a period of time of 6 hours, the dynamic evolution of the insuline and glucose concentrations are showed for the perturbations and non perturbations model.