Investigating sparse model for Continuous Glucose Monitoring in Type 2 Diabetes

Abstract

Continuous Glucose Monitoring (CGM) is a cutting-edge method for monitoring blood glucose levels at predetermined intervals. Type 2 diabetes is a long-term chronic lifestyle disease brought on by high blood sugar levels. In our investigation, we will use isolated liquid meals and CGM data to predict the glucose level inspiring from a widely used fitting method that is described in this work. Owing to the CGM data’s signifi-cant nonlinearity and the complexity of adequately modelling it, we restricted the use of non-linear approaches to a single time period during the day. After rigorous research, we discovered that we can draw inspiration from the underdamped case of a damped harmonic oscillator to simulate our data accurately. We then used the same technique to apply and model CGM data available for five patients (2 non-diabetic, 2 diabetic and one pre-diabetic). We have carefully analyzed and shown the data, day by day as well, so that it may be used for more comprehensive analyses of the glucose dynamics and the prescription of a diet or medication for diabetic patients since it demonstrates how glucose varies with simple liquid meals. Keywords: Non-Linear modelling, Data visualization and analysis, Continuous Glucose Monitoring, Harmonic Oscillator