The aim of this article is to describe how NN have been applied for the simulation of glucose—insulin metabolism, and to present two NN based personalized models for children with T1DM. The models, which are able to make short-term glucose predictions, are based on the combined use of MMs and NNs. The models are comparatively assessed using data about glucose levels, insulin intake, and diet during previous time periods, from four children with T1DM.
Key Terms in this Chapter
Continuous Glucose Monitoring Systems: Systems that monitor glucose levels at all times.
Type 1 Diabetes Mellitus: A condition characterized by disordered metabolism and inappropriately high blood sugar resulting from either low levels of the hormone insulin, or from abnormal resistance to insulin’s effects, coupled with inadequate levels of insulin secretion to compensate.
Personalized Models: Models that are individualised.
Artificial Neural Networks: Neural networks that include artificial intelligence.
Compartmental Models: Models that have distinct sections.