论文部分内容阅读
目的:通过对2003年1月～2008年12月重庆市法定报告传染病逐月发病率数据的分析,研究其变化规律,建立预测与监测的ARIMA时间序列模型。方法:用Box-Ljung统计量评价ARIMA模型的拟合度,用平均预测相对误差作为预测效果的评价指标。结果:重庆市法定报告传染病发病以年为周期,1年中4～6月为高发月,尤其是5月和6月最为严重。ARIMA(0,1,0)(0,1,1)12模型是重庆市法定报告传染病拟合的最佳模型,其拟合残差的方差为12.23,外推预测的平均相对误差为8.3%。结论:对传染病发病率历史数据进行时间序列分析是用于传染病监测的一个重要的内容。本研究所建立的ARIMA模型适用于重庆市传染病发病率预测与监测。
OBJECTIVE: To analyze the monthly incidence of infectious diseases reported by statutory reporting from January 2003 to December 2008 in Chongqing, and to study the variation law of ARIMA time series model of forecasting and monitoring. Methods: The Box-Ljung statistic was used to evaluate the fitting degree of ARIMA model, and the relative error of average prediction was used as the evaluation index of predictive effect. Results: The incidence of infectious diseases in statutory reporting in Chongqing Municipality was year-cycle. From April to June in 1 year, it was the highest incidence, especially in May and June. The ARIMA (0,1,0) (0,1,1) 12 model is the best model for the reporting of notifiable infectious diseases in Chongqing. The variance of the fitted residuals is 12.23 and the average relative error of the extrapolated predictions is 8.3 %. Conclusion: Time series analysis of historical data on the incidence of infectious diseases is an important part of the monitoring of infectious diseases. The ARIMA model established in this study is suitable for the prediction and monitoring of the incidence of infectious diseases in Chongqing.