Data-driven modeling and prediction of blood glucose dynamics: Machine learning applications in type 1 diabetes
作者：
Highlights：
• Numerous factors affect BG dynamics and an ideal predictor is expected to incorporate as much information as possible.
• Complexity of blood glucose dynamics pose a challenge to achieve accurate prediction in every scenario.
• Lack of proper formulation regarding penalty for errors in hypo/eug/hyper/ glycaemia regions.
• Challenges associated with errors incurred due to manual entry of diet and effect of CGM time lags are not well covered.
• Lack of a universal approach to quantify the approximate effect of physical activities, stress, and infection incidence.
摘要
•Numerous factors affect BG dynamics and an ideal predictor is expected to incorporate as much information as possible.•Complexity of blood glucose dynamics pose a challenge to achieve accurate prediction in every scenario.•Lack of proper formulation regarding penalty for errors in hypo/eug/hyper/ glycaemia regions.•Challenges associated with errors incurred due to manual entry of diet and effect of CGM time lags are not well covered.•Lack of a universal approach to quantify the approximate effect of physical activities, stress, and infection incidence.
论文关键词：Type 1 diabetes,Blood glucose dynamics,Blood glucose level prediction,Machine learning
论文评审过程：Received 26 April 2018, Revised 22 August 2018, Accepted 19 July 2019, Available online 26 July 2019, Version of Record 11 September 2019.
论文官网地址：https://doi.org/10.1016/j.artmed.2019.07.007
