Design of a Real-Time Physical Activity Detection and Classification Framework for Individuals With Type 1 Diabetes
Authors' conclusion
The proposed module showed high performance in detection and classification of exercise in real-time within a minute of exercise onset. Integration of this module into insulin therapy decisions can help facilitate glucose management around physical activity.
Study context
- Population
- Adults with T1D who performed structured sessions of aerobic, resistance or high-intensity interval exercise at varying times of day
- Participants
- 19 adults
- Outcome
- Exercise detection and classification performance
Methodological detailsrandom forest models, Inputs: tri-axial accelerometer, heart rate and CGM records
- Random forest models
- Inputs: tri-axial accelerometer, heart rate and CGM records
Performance
As reported in the abstract
- Exercise onset predicted within 1 minute, average accuracy
- 81%
- Exercise completion predicted within 1 minute, average accuracy
- 78%
- Activity type and intensity identified within (minutes from exercise onset)
- 2.38
- Activity type classification accuracy, announced exercise (test set)
- 74%
- Intensity classification accuracy, announced exercise (test set)
- 73%
- Activity type classification accuracy, unannounced exercise
- 65%
- Intensity classification accuracy, unannounced exercise
- 70%
Original publication
Design of a Real-Time Physical Activity Detection and Classification Framework for Individuals With Type 1 Diabetes. J Diabetes Sci Technol. Published 17 February 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →