ARMA模型在胆结石病发病率预测中的应用

目的探讨应用时间序列自回归移动平均模型（ARMA）预测胆结石病发病率的可行性。方法应用EV iews3.1软件对青海海西自治州2001~2006年逐月发病率进行ARMA建模拟合;按照残差不相关和简洁的原则确定模型结构,依据赤池信息准则（AIC）与施瓦茨准则（SC）、拟合优度确定模型的阶数。对所得模型的残差进行统计验证,依据残差序列图中实际值与拟合值的拟合效果确定最终模型。利用最终模型预测2007年的胆结石病月发病率,比较预测值和实际值,检验预测效果。结果AR IMA（1,1,1）模型能够较好地拟合及预测胆结石病月发病率值,除个别观测数据与模型预测数据相差较大外,模型所得结果与实际值非常接近。结论ARMA模型可以较好地模拟胆结石病发病率在时间序列上的变动趋势,可用于预测未来的胆结石病发病率趋势,是一种短期预测精度较高的预测模型。

Application of the ARMA model in forecast of the incidence of cholelithiasis

Objective To explore the application of time series anto-regression moving average（ARMA） model in prediction of cholelithiasis incidence.Methods EViews3.1 software was used to construct the ARMA model based on the monthly cholelithiasis incidence in Haixi Mongol Tibetan and Kazak Autonomous District of Qinghai province,from Jan 2001 to Dec 2006 with consideration of residual un-correlation and concision.Akaike information criterion（AIC）,Schwarz criterion（SC） and goodness of fit were used to determine the degree of the model.The constructed model was then applied to predict the monthly cholelithiasis incidence in 2007 and the incidence from the ARMA model was compared with the actual incidence so as to evaluate the model′s validity.Results（ARIMA）（1,1,1） model could better fit and predict the value of monthly incidence of cholelithiasis,except for some observed data.The prediction of the model was very close to the actual value.Conclusion The model of ARMA seems to fit exactly the changes in cholelithiasis incidence and to predict the future trend of incidence with a high prediction precision of short term time series.