Synthetic Continuous Glucose Monitoring (CGM) Signals

Published: 15-04-2021| Version 1 | DOI: 10.17632/chd8hx65r4.1
Simon Cichosz,
Alexander Arndt Pasgaard Xylander


Based on a trained Conditional Generative Adversarial Network (CGAN), the dataset contains 40,000 CGM days with a sampling frequency 288/day, equivalent to 940,000 hours of synthetic CGM. The dataset contains both signal resembling people with type 1 diabetes and healthy individuals. Profiles are categorized into four groups resembling different HbA1c levels: (1) below 6.5% (healthy without diabetes), (2) between 6.5% to <7%, (3) between 7% to <8% and (4) above 8%.