Individualized Prediction of Blood Glucose Outcomes Using Compositional Data Analysis

Mathematics(2023)

引用 0|浏览1
暂无评分
摘要
This paper presents an individualized multiple linear regression model based on compositional data where we predict the mean and coefficient of variation of blood glucose in individuals with type 1 diabetes for the long-term (2 and 4 h). From these predictions, we estimate the minimum and maximum glucose values to provide future glycemic status. The proposed methodology has been validated using a dataset of 226 real adult patients with type 1 diabetes (Replace BG (NCT02258373)). The obtained results show a median balanced accuracy and sensitivity of over 90% and 80%, respectively. A information system has been implemented and validated to update patients on their glycemic status and associated risks for the next few hours.
更多
查看译文
关键词
compositional data,decision support system,diabetes type 1,blood glucose prediction
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要