Research on Diabetes Classification based on R-R Interval and Main Weight Amplitude Ratio Percussion Entropy Algorithm

With the development of Internet technology, data processing, disease prediction and diagnosis based on computational algorithms have become hot issues. At present, computer algorithms have been widely used in the classification and prediction of diabetes, and have shown great application value in clinical aspects, the aim of this study was to use a noninvasive approach to simultaneously extract the main weight amplitude ratio sequence of the pulse signal (PPG) and the RRI sequence of ECG, and the coincidence rate between them was calculated by percussion entropy algorithm for the classification of diabetes. In this experiment, the synchronous principal weight amplitude ratio sequences and RRI sequences of healthy subjects (Group 1: 30 people) and diabetic subjects (Group 2: 28 people) were extracted, and the anastomoses between them were calculated using PEn. The normal physiological parameters of the human body and PEn were tested using an independent-sample T-test and Pearson correlation analysis. Finally, the multi-scale cross-approximate entropy (MCAE) was compared with PEn. The results show that $\operatorname{PEn}(\mathbf{p}\lt 0.001)$ distinguished diabetic patients from healthy individuals in a way that was both considerably distinct and superior to MCAE ($\lt 0.05$), and PEn was negatively correlated with both glycosylated hemoglobin ($r=-0.451, p\lt 0.001$) and fasting blood glucose ($r=-0.359, p\lt 0.001$). Not only do the results of this study indicate that the percussion entropy of subject weight ratio sequence and RRI sequence is superior in predicting the vascular health status of diabetes, but they also indicate that the principal weight amplitude ratio sequence and RRI sequence can be used as characterization parameters for diabetes, demonstrating certain research significance for the early screening of diabetes.