摘要
目的:探讨应用ARIMA模型预测宝安区某街道其它感染性腹泻发病率的可行性。方法:应用SPSSl3.0软件对2005年～2009年宝安区某街道其它感染性腹泻逐月发病率进行ARIMA模型建模拟合,用所得到的模型对2010年各月发病率进行预测,并评价其预测效果。结果:宝安区桌街道其它感染性腹泻发病率每年11月为发病高峰,ARIMA(0,1,1)(0,1,0)12模型是其拟合的最佳模型,其预测结果和实际值绝对误差的绝对值最大为930.47,最小为1.96,平均值214.83,平均相对误差百分比39.04%。结论:模型虽然起到一定的预测效果,但预测精度仍存在误差,可通过积累新的周期数据对ARIMA模型进行修正和重新拟合,也可尝试新的预测方法或其他模型,才能加强和保证预测的精度。
Objective： To probe the feasibility of the incidences of other infectious diarrheas at a township,Bao＇an district predicted by ARIMA model. Methods： Using SPSS13.0 to build the ARIMA model and to fit the month-by-month incidences of other infectious diarrheas from 2005 to 2009 at a township,Bao＇an district, to predict the incidences from January to December,2010, and to evaluate the prediction effects. Results： November is the incidence peak of other infectious diarrheas at a township,Bao＇an district. The best fitting model is ARIMA （0,1,1）（0,1,0）12 model.Absolute errors between prediction results and actual value , the maximum is 930.47,the minimum is 1.96,the average is 214.83,the percent of average relative error is 39.04%. Conclusions： The ARIMA model has some prediction effects, but there is error to the accuracy. Revising and re-fitting the ARIMA model by accumulating new period data or trying new prediction ways and other models can strengthen and guarantee the accuracy.
出处
《现代生物医学进展》
CAS
2011年第16期3138-3142,共5页
Progress in Modern Biomedicine
基金
宝安区科技局2009年基金项目(2009417)
关键词
ARIMA模型
感染性腹泻
发病率
预测
Model ofautoregressive integrated moving average
Infectious diarrhea
Incidence
Prediction
