时间序列分析在洞庭湖区双退试点血吸虫病发病预测中的应用

Application of time series analysis in the prediction of schistosomiasis prevalence in the areas of “breaking dikes or opening sluice for waterstore” in Dongting Lake

目的 :通过比较时间序列分析中指数平滑法、移动平均法、自回归分析及自回归综合移动平均法 (Autoregressiveintegratedmovingaveragemodel,ARIMAmodel)在洞庭湖区退田还湖试点集成垸 1 990 / 2 0 0 2年血吸虫病患病率预测中的优劣 ,为当地退田还湖试点的血吸虫病发病找到一个较适合的预测模型 ,为防治工作提供参考依据 .方法 :用时间序列分析各方法建模预测 ,比较各方法的拟合优度、误差平方和及预测值的相对误差 ,确定最佳预测方法 .结果 :指数平滑法、移动平均法、自相关分析及ARIMA法的 1 993/ 2 0 0 2年患病率预测值年平均相对误差 (% )和误差平方和ARIMA模型最小 .结论 :集成垸 1 990 / 2 0 0 2年患病率预测中 。

AIM: To provide the fittest model for prediction of schistosomiasis prevalence in Jicheng of “breaking dikes or opening sluice for waterstore” in Dongting Lake by comparing the results of moving average, exponential smoothing, autoregressive model and autoregressive integrated moving average model (ARIMA model ) from 1990 to 2002, to provide reference to schistosomiasis preventive work. METHODS: The fittest model was chosen by comparing the goodness of fit, error sum of square and relative error of four statistical methods. RESULTS: The average predicted relative errors and error sum of square of ARIMA from 1993 to 2002 are the smallest. CONCLUSION: The fittest model in the prediction of schistosomiasis prevalence in Jicheng of “breaking dikes or opening sluice for waterstore” in Dongting Lake from 1990 to 2002 is ARIMA model.