时间序列模型在肾综合征出血热发病率预测中的应用

A time series model in incidence forecasting of hemorrhagic fever with renal syndrome

目的探讨时间序列模型预测肾综合征出血热（HFRS）发病率的适用性。方法应用临沂市1982～1999年HFRS月发病率资料拟合HFRS月发病率预测模型。结果利用时间序列模型中的ARIMA模型预测山东省临沂市2000～2002年3年逐月发病率，2000年预测值的误差最小。结论ARIMA模型可用于预测HFRS月发病率，其短期预测精度较高。

Objective To discuss if it is possible to forecast hemorrhagic fever with renal syndrome （HFRS） incidence with time series predictive models. Methods The HFRS incidence model was fit with time series models based on the data from 1982 to 1999 in Linyi City. The monthly incidence in Linyi City from 2000 to 2002 was predicted with ARIMA models. Results The data of 2000 had the closest tolerance. Conclusion The ARIMA model can be used to forecast HFRS incidence with high predictive precision in the short-term.