Analysis and Prediction of the Prevalence of Asthma in Mainland China Based on Back Propagation Neural Network

Asthma is part of the common chronic diseases in China and a serious public health problem. The data in this article came from the China Health and Retirement Longitudinal Study (CHARLS) database and the China Meteorological Data Service Center, and the correlation between urban asthma prevalence and meteorological factors and population characteristics was analyzed. The correlation coefficient between the prevalence rate of urban asthma and the standard deviation of monthly mean air pressure was 0.271, and the correlation coefficient for the proportion of people who had lung disease was 0.609. Select variables with significant correlation and apply the Back Propagation Neural Network (BPNN) model to analyze and foresee the prevalence of asthma in the surveyed cities. Among them, the prediction result of 30% of the prediction data is mean absolute error MAE=1.24, and the determination coefficient R2=0.674, which has a decent prediction effect.