Prediction Of Adverse Glycemic Events From Continuous Glucose Monitoring Signals By Gradient Boosting Algorithm
In Diabetes therapy,the most important step includes maintaining the glucose level in the proper euglycemic range.CGM devices have been introduced which helps to see the glucose levels and helps in maintaining in the desired range. The challenge here is detecting the glucose level and predicting the critical levels in the glucose levels with the collected data.This study aims to ﬁll this gap, by carrying out a comparative analysis among the most common methods for glucose event prediction. Both regression and classiﬁcation algorithms have been implemented and analyzed, including static and dynamic training approaches. The dataset consists of 89 CGM time series measuring India diabetic subjects for 7 subsequent days. Performance metrics, speciﬁcally deﬁned to assess and compare the event prediction capabilities of the methods, have been introduced and analyzed. The numerical results show that a static training approach exhibits better performance, in particular when regression methods are considered. However, classiﬁers show some improvement when trained for a speciﬁc event category, such as hyperglycemia, achieving performance comparable to the regressors, with the advantage of predicting the events sooner.