五种预测方法在退田还湖区血吸虫病发病拟合效果评价.docVIP

五种预测方法在退田还湖区血吸虫病发病拟合效果评价 　 作者：赛晓勇，邢秦菊，孟定茹,贾玉然，蔡凯平，李岳生，周晓农 【关键词】 统计预测；ARIMA模型;血吸虫病；退田还湖 　　Comparison of predicting effect of schistosomiasis prevalence by 5 statistical models in the areas of “breaking dikes or opening sluice for water store” in Dongting Lake 　　【Abstract】 AIM: To compare the predicting effect of schistosomiasis prevalence by 5 different statistical models including Moving Average, Exponential Smoothing, Autoregressive Model, Autoregressive integrated moving average model (ARIMA Model) and Grey Model in the areas of “breaking dikes or opening sluice for water store” in Dongting Lake and to provide a fitted model for local schistosomiasis preventive department. METHODS: The 5 different statistical models were applied to predict the schistosomiasis prevalence in some experimental sites and Error Sum of Square (ESS), Average Relative Errors (ARE), Average Errors (AR) of 5 models were compared. RESULTS: ESS, ARE and AR of Grey Model in Jicheng were smallest; ESS and AR of ARIMA Model in Haohou were smallest; ARE of Autoregressive Model was smallest. CONCLUSION: Different models fit different places. The predicting effects of Grey Model and ARIMA Model are best among the 5 models. 　　【Keywords】 Statistical prediction；ARIMA Model; Schistosomiasis；Breaking dikes or opening sluice for water store 　　【摘要】 目的： 比较移动平均法、指数平滑法、自回归法、ARIMA法和灰色预测法在退田还湖地区试点血吸虫病发病拟合效果的优劣，为当地血防部门提供较为适合的拟合方法. 方法： 应用五种方法对集成垸试点和濠口试点血吸虫病患病率建模预测并比较拟合值的绝对误差、相对误差和误差平方和. 结果： 集成垸试点灰色预测法拟合值的平均绝对误差、平均相对误差和误差平方和最小；濠口试点平均绝对误差、误差平方和以ARIMA法最小，平均相对误差以自回归法最小. 结论： 不同的拟合模型适用于不同的试点；两试点以灰色预测和ARIMA模型拟合效果较好. 　　【关键词】 统计预测；ARIMA模型;血吸虫病；退田还湖 　　0引言 　　1998年我国开始退田还湖,使血吸虫病中间宿主钉螺孳生环境发生了变化. 在应用不同方法对血吸虫病病情预测研究的基础上，对移动平均法、指数平滑法、自回归法、ARIMA法和灰色预测法进行了比较与评价，为退田还湖区血防部门找到相对精确的定量拟合方法. 　　1材料和方法 　　1.1材料 　　收集退田还湖地区华容县的集成垸试点（双退点，即退人又退田,该垸1998年完全废弃用于泄洪）和濠口试点（单退点，退人不退田即洪水期人转移、洪水过后返回种田）1990~2003年连续粪检阳性率的病情资料. 集成垸试点退田还湖后滞留人口2600人，面积为2200万平方米；濠口试点常住人口1176人，面积为297万平方米，均为湖南省血吸虫病重灾区监测试点. 全部病情资料由湖南省血防所及华容县洪山头镇血防站和澧县小渡口血防站提供. 　　1.2方法 　　1.2.1移