Ahead of time prediction of nocturnal hypoglycemic events from Continuous Glucose Monitoring data in people with type I diabetes by Machine Learning-based approaches

Type 1 diabetes can lead to short and long term complications due to a bad control of the blood glucose level inside safe limits, called euglycemic range. Continuous Glucose Monitoring (CGM) systems allow diabetic people having a better management of the glycemia and encourage the development of predction algorithms. This thesis work aimed to develop a method to predict nocturnal hypoglycemic events using only the CGM data collected in the previous day exploiting machine-learning approaches.