Datasets
Standard Dataset
Identification and validation input-output diabetic dataset (in-silico)
- Citation Author(s):
-
MartinDodek
Faculty of Electrical Engineering and Information Technology Slovak University of Technology in Bratislava - Submitted by:
- Martin Dodek
- Last updated:
- Thu, 08/11/2022 - 11:59
- DOI:
- 10.21227/qszf-v124
- Data Format:
- License:
270 Views- Categories:
- Keywords:
Abstract
The glycemia response for this experiment was obtained in-silico, using the complex physiology-based nonlinear simulation model with the basal glycemia 6 mmol/l and the corresponding basal insulin administration rate 0.01 U/min.
The data acquisition experiment was designed to mimic the regular insulin treatment of a type 1 diabetic subject during the 6-day period with an overall number of 25 meals and a total carbohydrate amount of 433 g.
The virtual continuous glucose monitoring readings were sampled with the sample time 20 min.
The glycemia measurements were distorted by the additive white noise with the standard deviation 0.1 mmol/l.
The insulin treatment was executed according to the modified bolus calculator with the parameters adjusted as IS=15 mmol/l/U, ICR=8.0 g/U.
There are two independent files - the identification and the validation dataset.
Each comprise three Matlab variables:
- Glycemia readings [mmol/l]
- Insulin administration rate [U/min]
- Carbohydrate intake rate [g/min]
Dataset Files
- Identification dataset ident_data.mat (4.68 kB)
- Validation dataset valid_data.mat (4.67 kB)
Comments
.
.