The Relationship Between QT Interval, Dispersion Reference Value And Geographical Factors In Chinese Adult ECG Correction

Corrected QT interval (QTc) and corrected QT dispersion (QTcd) are important indexes of clinical electrocardiogram (ECG). QT interval and QT interval dispersion are affected by many factors, such as heart rate, age, sex, drugs and autonomic nerves.The geographical environment also has impacts on the reference value of ECG.This paper selects adult electrocardiogram QTc and QTcd reference value to explore the mechanism how geographic factors affect medical reference value.The study is focus on the relationship between reference value of QTc, QTcd and 25 geographical factors.selecting 34505 samples from China.On the basis of correlation analysis, derived the correlative geographical factors from the database. For analyzing the reference value and the geographical factors'complicated relationship better, this paper establishes support vector regression machine model, curve estimation model and ridge regression model.Then build up the combination forecast model, mapped the geographical distribution of different gender and discuss the influence mechanism.Through the study, we got the following conclusions:(1)Based on the correlation analysis, there is a significant relationship between the two reference values and geographical factors.5 geographical factors had significant correlations with male and female QTc value.7 geographical factors had significant correlations with male and female QTcd value.(2)The simple weighting prediction model is chosen as the best model to predict the male QTc reference value. The weight of ridge regression model is 1/2, the weight of curve estimation is 1/6, and the weight of support vector machine model is 1/3.The optimal weighting prediction model is chosen as the best model to predict the female QTc reference value. The weight of ridge regression model is 0.07, and the weight of support vector machine model is 0.93.The optimal weighting prediction model is chosen as the best model to predict the male QTcd reference value. The weight of ridge regression model is 0.973, and the weight of curve estimation is 0.027.The standard deviation weighting prediction model is chosen as the best model to predict the female QTcd reference value. The weight of ridge regression model is 0.392, the weight of curve estimation is 0.411, and the weight of support vector machine model is 0.197.(3)The distribution map of QTc and QTcd reference values in healthy Chinese adults successful shows its visual renderings. It is suitable to discover the rules of space, and also facilitate the clinical medical workers to directly read local reference from the map.