Supervisor: Giacomo Cappon
Co-supervisor: Luca Cossu
Co-supervisor Department/Company: DEI
Creation Date: 08/07/2025 13:35
This work proposes a data-driven approach to model the metabolic trajectory of patients with Type 1 Diabetes by integrating Digital Twin technology and Markov models. Using a dataset of 14,888 patient records, I| first clustered the data to identify distinct metabolic profiles. From these clusters, I derived Markov models to capture probabilistic transitions between metabolic states. The models were then used to generate synthetic daily trajectories, which were simulated in ReplayBG.
Dataset type: Already acquired data
Dataset description: 14,888 daily records of people with type 1 diabetes including continuous glucose monitoring, insulin injections, and meal intakes.
List of Methods: Digital Twins Markov Models Clustering Techniques
Analisi dei Dati Biologici Machine Learning