Study on the medical meteorological forecast of the number of hypertension inpatient based on SVR

3rd International Conference on Advances in Energy, Environment and Chemical Engineering

能源学;生态环境科学;化学工业

Zhai, Guangyu^1,2 ; Chai, Guorong^1 ; Zhang, Haifeng^1

School of Management, Lanzhou University, Lanzhou

730000, China^1

School of Economics Management, Lanzhou University of Technology, Lanzhou

730050, China^2

关键词: Cross validation algorithms;    Dependent variables;    Independent samples;    Independent variables;    Meteorological factors;    Neural network predictions;    Optimal predictions;    Support vector regression (SVR);   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/69/1/012059/pdf
DOI  :  10.1088/1755-1315/69/1/012059

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

The purpose of this study is to build a hypertension prediction model by discussing the meteorological factors for hypertension incidence. The research method is selecting the standard data of relative humidity, air temperature, visibility, wind speed and air pressure of Lanzhou from 2010 to 2012(calculating the maximum, minimum and average value with 5 days as a unit ) as the input variables of Support Vector Regression(SVR) and the standard data of hypertension incidence of the same period as the output dependent variables to obtain the optimal prediction parameters by cross validation algorithm, then by SVR algorithm learning and training, a SVR forecast model for hypertension incidence is built. The result shows that the hypertension prediction model is composed of 15 input independent variables, the training accuracy is 0.005, the final error is 0.0026389. The forecast accuracy based on SVR model is 97.1429%, which is higher than statistical forecast equation and neural network prediction method. It is concluded that SVR model provides a new method for hypertension prediction with its simple calculation, small error as well as higher historical sample fitting and Independent sample forecast capability.