Relationship Between E / A Reference Value Of Mitral Blood Flow Spectrum And Geographical Environment In Chinese Adults

Heart disease as a high incidence of disease in our country, it always had been the focus of medical research, the reference values of cardiac function can effectively help us to strengthen the accuracy of the heart disease prognosis. Mitral valve flow spectrum E/A as an important indicator of heart function, the reference value of it has a significant meaning for the clinical diagnosis. The physiological function of it has a certain influence on it, but the relationship between the natural geographical environment is also worth exploring, from the perspective of geographical influence, we try to explore the relationship between the geographical environment and mitral valve flow spectrum E/A in this study.Through the method of combining the literature searching and investigation, we collected mitral valve flow spectrum E/A reference values of 25451 healthy people from 93 cities in China, we would build the prediction model based on the known data to predict the reference values of other unknown locations.We used the Moran index to determine the spatial relationship, selected 18 geographical factors, examined the significance between E/A and geographical factors by correlation analysis, through the significance test, and extracted 9 significant factors to build 4 model, they are principal component model, the ridge regression model, the artificial neural network model, the support vector machine model, these models have their advantages and disadvantages respectively. In order to extract more useful information and improve the prediction precision of the model, we choose the 4 kinds of single model based on weighted synthesis to build 3 kinds of combination forecast model, through accuracy comparison to choose the best model. By using the spatial analysis to get the trend of reference value and geographic distribution.Through the study, we got the following conclusions:(1) The spatial autocorrelation of the reference value of E/A was significant. E/A was affected by the space geographical environment, that is to say, different environment could make different geographical distributions of E/A.(2) Based on the correlation analysis,9 geographical factors had significant correlations with E/A, they were latitude, annual sunshine duration, annual mean air temperature, annual mean relative humidity, annual precipitation amount, annual range of air temperature, topsoil sand fraction, topsoil silt fraction, topsoil bulk density, which extracted the 9 geographical factors to build the model.(3) The optimal weighting model was chosen as the best model to predict the reference values of E/A. The weight of principal component model is 0.073, the weight of ridge regression model is 0.217, the weight of artificial neural network model is 0.014,the weight of support vector machine model is 0.696.This selection is based on the error evaluation between these models, the forecast model could synthesize the advantages and disadvantages of these models, and extracted more effective prediction information, improved the prediction precision.(4) The spatial distribution trend of E/A is high in southwest, low in northeast. From west to east, the value of E/A increased first and then decreased. From south to north, the value of E/A decreased. High value and low value have obviously spatial distribution. The low value concentrated in the eastern part of Xinjiang, Inner Mongolia, and part of Heilongjiang, Liaoning and Jilin province, the high value concentrated in the part of Sichuan, Chongqing, Yunnan, and the southeast of Tibet.