基于公开数据的SARS流行规律的建模及预报

Models and Prediction of the SARS' Prevalence

基于全球与我国公开发表的有关SARS确诊病例、疑似病例和死亡病例的实际数据,建立有关流行病学的模型是本研究的目的。我们探讨了SARS的简单流行病学微分方程模型,建立了北京医院尚有确诊病例的时间序列的ARMA模型,得出6月底左右北京医院SARS病人数将会降到60人以下。同时,通过经验模型和贝叶斯方法对病死率进行了初步估计。

Building some epidemiology models based on the reported data abuout SARS' confirmed cases, suspected cases and the number of deaths etc. is what this essay mainly focused in. We discussed the simple differential equations in epidemiology, built the time series' ARMA model to predict the number of patients in Beijing's hospital, and get the prediction that there will be no more than 60 patients in Beijing's hospital until the end of June which will satisfies WHO. At the same time we got some preliminary estimation of the fatality rate through empirical model and the Bayes method.