Type 1 Diabetes Mellitus Prediction Model Based on Forecasting Algorithm

Forecasting, as an approach, performs an essential role in our life nowadays. Recently this technique has pervaded the healthcare field, where predictions can be simulated to prevent climacteric stages for the patient. This paper deals with T1D illness considered a chronic condition that constantly threatens the life of patients between the age of 4-14. The main problem of this illness is related to the lack of insulin in the blood. As the human body is designed as a black box, this research aims to understand the functioning of the body influencing the blood glucose value. The aim of this work is to build a system that predicts the change in blood glucose value in the future; the patients will be ready to schedule the different activities in their life and sufficiently handle their T1D. The system is based on machine learning models to forecast the future and maintain the glucose value in a safe range, avoiding critical cases. The data are from a 6-year-old child, where CNN, RNN and LSTM models are considered in this paper to make predictions; the obtained results show a good accuracy; meanwhile, other parameters could be added to have more promising results.