SARIMA模型用于新疆乙肝发病率预测的探讨

摘要:目的　探讨SARIMA模型用于新疆乙肝月发病率预测的可行性，为制定防控策略提供科学参考。方法　采用带有季节性的自回归移动平均模型（SARIMA），分析将其用于新疆乙肝发病率预测的可行性，对模型进行参数估计及残差检验，根据AIC及BIC准则确定最适合的SARIMA模型，讨论该模型的拟合及预测效果。结果　SARIMA（1，1，2）（0，1，0）12模型很好地拟合了新疆乙肝月发病率变化规律，模型预测值与实际值间的相对误差及均方误差较小。结论　SARIMA（1，1，2）（0，1，0）12模型能够较好地用于新疆乙肝月发病率预测，可为新疆的乙肝防控提供科学参考。

SARIMA model for the prediction of hepatitis B incidence in Xinjiang

Abstract: Objective The study was aimed to discuss the feasibility of using seasonal autoregressive integrated moving average (SARIMA) model to predict the incidence of hepatitis B in Xinjiang, in order to provide scientific reference for formulating preventive and controlling strategies. Methods The feasibility of using SARIMA to forecast the incidence of hepatitis B was discussed. The parameter and doing residual test of SARIMA model was estimated according to Akaike information criterion (AIC) and Schwarz criterion (BIC) to confirm the best SARIMA model from possible models. Then the effect of fitting and prediction of the best SARIMA model was analyzed. Results The SARIMA was able to fit the trend of hepatitis B incidence well. The average relative error and root mean square error between actual value and prediction value were small. Conclusion The SARIMA can be used to predict hepatitis B incidence in Xinjiang, which can provide scientific reference for formulating prevention and control strategies.