DEVELOPMENT OF A NEURAL NETWORK MODEL ARCHITECTURE FOR PERSONALIZED INSULIN DOSAGE CALCULATION IN TYPE I DIABETES
Ghenadie USIC, Aurelia PROFIR, Moldova State University
Abstract
Type I diabetes is a chronic autoimmune disease characterized by insulin hormone deficiency in the body, requiring patients to rely on external administration of this hormone for survival. Additionally, blood glucose monitoring
and precise insulin dose calculation are essential aspects in managing this disease. The current research proposes an
approach that uses machine learning techniques to optimize and personalize insulin dose calculations for patients
with type I diabetes. The developed model is based on neural networks and has been trained using pharmacodynamic
profiles and a simulated dataset. The research was conducted over a period of two months, focusing on adapting the
algorithm to individual patient characteristics to improve blood glucose level management.
Keywords: artificial neural network, machine learning, type 1 diabetes management, predictive model, glycemic
control, insulin dosage, real-time analysi.
