Prediction of Daytime Hypoglycemic Events Using Continuous Glucose Monitoring Data and Classification Technique

-- Daytime hypoglycemia should be accurately predicted to achieve normo glycemia and to avoid disastrous situations . Hypoglycemia, an abn ormally low blood glucose level, is divided into daytime hypogly cemia and nocturnal hypoglycemia . In this paper, we propose new predictor variables to predict daytime hypoglycemia using continuous glucose monitoring (CGM) data. We apply classification and regression tree (CART) as a prediction method . The evaluation results showed that our model wa s able to detect almost 80% of hypoglycemic events 15 min in advance, which was higher than the existing methods with similar conditions . T he proposed method might achieve a real - tim e prediction as well as can be e mbedded into BG monitoring device. Diabetes is one of the most common chronic diseases in the world, affecting 2.72 million individuals (10% of the population) in the Korea [1] and 29.1 million individuals (9.3% of the populat ion) in the USA with increasing incidence [2] . Diabetes can be th e cause of kidney failure, lower - limb amputations, and blindness among adults [2] . A chievement of excellent glycemia is most important task to diabetic patients in both type 1 and type 2 diabetes. D iabetic patient s should maintain euglycemic blood glucose (BG) levels while all day and be required the wisdom to avoid hyper - and hyp oglycemia [3] . Especially, the patients who treated w ith an insulin are at risk for developing hypoglycemia. Population - based data indicate that 30 - 40% o f people with type 1 diabetes ex perience an average of three episodes of severe hypoglycemia each year; those with insulin - treated type 2 diabetes experience about one episode of that each year. Also, individuals with type 1 diabetes experienced about 43 symptomatic (not only severe) episodes annually; insulin - treated individuals with type 2 diabetes experienced about 16 episodes annually [4] . The s ymptomatic hypoglycemic e pisode mean s that the patients feel the symptoms of s hakiness, sweating, hunger, irritability or headache [5] . H ypoglycemia is a significant challenge for a precise insulin therapy [6] .