A Novel Classification Indicator of Type 1 and Type 2 Diabetes in China

摘要: Because of the differences treatment, it is extremely important to classify types diabetes, especially for diagnosis made by clinician. In this study, we proposed a novel scheme calculating an indicator classifying which contains two stages: first model feature extraction, 17 features are automatically extracted from curve glucose concentration acquired continuous monitoring system (CGM); second diabetes parameter regression based on ensemble learning algorithm named double-Class AdaBoost. 1050 curves type 1 and 2 diabetics were at Department Endocrinology in People’s Hospital Zhengzhou University China, upper threshold μ was set 7 mmol/L, 8 mmol/L, 9 mmol/L, 10 mmo/L, 11 mmol/L respectively according guideline WHO. The experiments show that coincidence rate our clinical 90.3%. extends criteria diagnosing provides doctors with scalar 2.