基于运气理论探索香港呼吸道传染病与气象因素的相关性

Correlation between Meteorological Factors and Incidence of Respiratory Infections in Hong Kong Based on Yunqi Theory

目的:根据中医“三年化疫”的运气理论,建立反向传播算法(BP)神经网络的传染病预测模型,探索香港呼吸道传染病(respiratory infectious disease,RID)与气象因素之间的相关性。方法:根据香港卫生署卫生防护中心提供的1997—2019年RID的发病数据与香港天文台提供的1994—2019年的平均气温、相对湿度、降雨量等9个气象因素,利用BP人工神经网络的统计学方法,探索性构建“RID-运气-气象”预测模型。结果:香港地区RID在三之气(5—6月份)高发;利用2年前至当年(共3年)的气象因素建立的RID发病预测模型效果最佳,其中的关键因素是平均水汽压。结论:香港RID发病与前期(如:1~2年前)的气象因素存在一定的相关性,这在一定程度上证明了“三年化疫”学说的客观性,可能为香港今后的RID预警发挥作用。

Objective:To establish a respiratory infectious disease(RID)prediction model of BP neural network based on the Yunqi theory of'pestilence occurring after 3 years',and to explore the correlation between RID and meteorological factors in Hong Kong.Methods:According to the RID data from Hong Kong Centre For Health during 1997~2019 and meteorological factors(average air temperature,relative humidity,rainfall,9 meteorological factors)from Hong Kong Observatory during 1994~2019,a prediction model was established through BP artificial neural network thus to explore the correlation between meteorological factors and RID in Hong Kong.Results:The incidence of RID was the highest in the 3rd qi(May~June).The prediction of RID was successfully established based on the meteorological factors from 2 years ago to the current year(three years in total).The important factor was average air pressure.Conclusion:There is a certain correlation between the incidence of RID in Hong Kong and the previous meteorological factors,which proves the objectivity of'pestilence occurring after 3 years'.This concept will help preventing and predicting RID in Hong Kong.