关键词: 气象条件, 下呼吸道疾病, 逐步回归分析, 人工神经网络, 预测模型

Abstract: :Respiratory diseases greatly affected children’s health, and its occurrence was related to meteorological conditions closely. In order to analyze the effects of meteorological conditions on children’s lower respiratory diseases in Qinhuangdao, predict the number of patients and provide new method for medical meteorological service, the data of children with lower respiratory diseases from 2015 to 2016 and the meteorological data within the same time were used, prediction models were established by stepwise regression analysis and BP artificial neutral network separately, and the prediction effects were evaluated. The results show that the meteorological conditions, especially the staged weather changes and climate anomalies, had a significant effect on the patients’ number with these diseases. The number of patients was negatively correlated with air temperature and average relative humidity, and positively correlated with air pressure, wind speed and temperature fluctuation in 72 hours, and good correlation was showed between patient number and air temperature, followed by air pressure and average relative humidity. The prediction accuracy of the stepwise regression model and BP artificial neural network model was 72.75% and 76.30%, respectively. Between the two models established, the overall performance of BP artificial neural network model was better.

Key words:  meteorological condition, lower respiratory tract infection, stepwise regression, BP artificial neural network, prediction model