ViBA

Data Augmentation via Digital Twins to Develop Personalized Deep Learning Glucose Prediction Algorithms for Type 1 Diabetes in Poor Data Context.

Prendin F, Facchinetti A, Cappon G. · IEEE transactions on bio-medical engineering · 2026 · Article

10.1109/tbme.2025.3635264

Verified evidence

Model evaluation

946eb155-c0df-5096-9aad-970dfef9dea2
Prediction or fitting error

Prediction source Version of record · PDF page(s) 8

Study b0cd63a0-98f1-577a-8bfb-38a0d3d4a093 · Prediction record (JSON)

Prediction fields

Concrete comparator
neural net trained on five days of original plus augmented records versus neural net trained on 40 days of original records · PDF page(s) 8
Endpoint
root_mean_square_error · PDF page(s) 8
Family
prediction_error · PDF page(s) 8
Measured variable
glucose_concentration · PDF page(s) 1, 8
Measurement role
prediction_error · PDF page(s) 8
Metric
root_mean_square_error · PDF page(s) 8
Model
machine_learning · PDF page(s) 1, 8
Population
human_observations · PDF page(s) 1, 8
Reported time window
30-minute glucose forecast horizon for the RMSE comparison · PDF page(s) 6
Role
prediction_error · PDF page(s) 8