运用时间序列模型对麻疹流行趋势的预测与分析

Application and analysis of ARIMA model on forecasting in the incidence rate of measles

目的:应用时间序列模型对江苏省麻疹疫情进行预测分析,并探讨提高预测实用性的思路。方法:以1980年～2005年江苏省麻疹发病资料建立时间序列分析模型,以2006年的发病资料作为模型预测效果的考核样本。先采用差分方法对序列资料进行平稳化,然后进行定阶并估计参数,建立ARIMA模型,最后对预测结果进行分析,并利用模型对2009年强化免疫效果进行简单评价,探讨对疫情进行预警的方法和思路。结果:江苏省麻疹的发病趋势自2005年明显上升之后保持平稳,但有小幅波动,这与实际情况吻合。结论:用时间序列模型对传染病发病情况的拟合结果满意,预测效果良好,可为麻疹的防治提供一定的科学依据。

Objectlve：To forecast the incidence rate of measles in Jiangsu province by ARIMA model and discuss the method to improve its veracity. Methods：Based on the reported data of measles of Jiangsu province from 1980 to 2005,model was fitted to check out the sample from 2006 to verify its practicability. Firstly,used the difference method to make the data sequence become placid. Secondly,parameters of model was estimated and set up a product season model by deciding the rank of it. Finally,the good- ness-of-fit was given to analyze and evaluate the model. By which to explore the method to early warning of measles. Results： After an obviously ascending in 2005, the measles incidence rate kept steady, but for some slight fluctuation, which was fit to the practical sit- uation. Conclusion：It is practical to apply the approach of ARIMA product season model to predict measles in Jiangsu Province by test, providing a scientific basis for the prevention and control of the epidemic.