两种季节性时间序列模型在细菌性痢疾发病率预测中的应用研究

Application of two kinds of seasonal time series models in forecasting bacillary dysentery incidence

目的评价两种季节时间序列模型对武汉市江汉区细菌性痢疾月发病率的预测效果,为选择适宜的预测方法提供科学参考依据。方法收集2005年1月至2013年12月武汉市江汉区菌痢月发病资料,分别运用季节指数趋势预测模型和SARIMA模型进行建模,选择较优模型对2014年1~6月菌痢月发病率资料进行预测和效果评价。结果两种模型预测值与实际值拟合趋势基本一致,但季节指数趋势模型拟合效果较好,预测2014年上半年疫情发展特点为1~3月呈缓慢上升趋势,4~6月呈快速上升趋势。结论季节指数趋势模型对江汉区菌痢月发病水平进行短期预测具有较高的预测精度。

Objective To evaluate the effects of two seasonal time series models in predicting the incidence of bacillary dysentery, and provide references for choosing a suitable prediction method. Methods The monthly incidence data of bacillary dysentery from January 2005 to December 2013 in Jianghan District were collected. The data was modeled with the seasonal index trend model and SARIMA model. The results were compared to find a better fitted model which was then used to forecast the data from January to June in 2014. The model was further evaluated by comparing the forecast with the actual incidence. Results Forecast data of bacillary dysentery from January 2005 to December 2013 of the two models were analyzed, and were consistent with real-world data. The seasonal index trend model had a better fitted effect. This model predicted that the incidence of bacillary dysentery of Jianghan District of Wuhan City from January to March 2014 would increase slightly, and rise rapidly from April to June. Conclusion Seasonal index trend model can be used as a short-term forecasting model to predict the monthly incidence rate of bacillary dysentery in Jianghan District , Wuhan, and had a higher prediction accuracy.