基于中医运气理论探讨流行性乙型脑炎高发与前期气象因素的相关性并建立预测模型

目的 依据中医运气学说，研究北京地区流行性乙型脑炎（以下简称"乙脑"）的高发与前期（当年~3年前）气象因素的相关性，并建立BP人工神经网络的医疗气象预测模型。方法 收集北京地区1970年~2004年35年的气象资料和乙脑发病资料，筛选出乙脑的高发时段，采用BP人工神经网络方法，分别从当年、1年前、2年前、3年前的气象情况进行分析，建立乙脑的气象预警模型。结果 乙脑的高发期主要集中在每年的四之气（P＜0.01）。利用前期气象因素皆可成功建立乙脑高发的预测模型，其预测精度均高于80%，其中贡献度最大的气象因素分别是当年初之气的平均风速、1年前三之气的平均相对湿度、2年前初之气的平均风速、3年前二之气的平均风速。结论 北京地区乙脑的高发与前期（当年~3年前）的气象因素具有相关性，尤其是利用3年前的气象因素建立的预测模型效果最佳。

Objective To study the correlation between high incidence of Japanese encephalitis and meteorological factors in the same year and three years ago based on TCM theory of five circuits and six qi, and establish a medical meteorological forecast model of BP artificial neural network. Methods The data of the monthly incidence of Japanese encephalitis and meteorological factors were collected from 1970 to 2004 (altogether 35 years) in Beijing, from which the period of high incidence of Japanese encephalitis was extracted. Then the forecast model of Japanese encephalitis was established by applying BP artificial neural network analysis to meteorological factors recorded in the same year, one year ago, two years ago and three years ago. Results Fourth qi was the peak period of outbreak of Japanese encephalitis (P < 0.01). All forecast models based on recorded meteorological factors yearly were successfully established with forecast accuracy higher than 80%. Additionally, the most important meteorological factors calculated were average wind speed in first qi in the same year, average relative humidity in third qi in the one year ago, average wind speed in first qi in the two years ago, and average wind speed in second qi in the three years ago. Conclusion The high incidence of Japanese encephalitis in Beijing is related to meteorological factors in the year and the three years ago. Furthermore, the forecast model established based on the meteorological factors of the three years ago is evaluated as the best one.