基于时间序列模型对甲型病毒性肝炎的预测研究

Prediction of viral hepatitis A based on time series model

目的探讨时间序列模型在甲肝发病预测的应用,为下一步采取防控措施提供科学依据。方法基于宜昌市2005-2015年逐月甲肝发病率建立两种模型,对2016年甲肝的发病率进行预测,并将预测值与实际值进行拟合评价。结果 ARIMA模型首先要求数据平稳,宜昌市的甲肝发病存在季节性波动,为不平稳序列,但2010年之后数据较为平稳,经对2010-2015年甲肝月发病率进行季节性差分、差分处理,新数列为平稳序列(游程检验法Z=1.447,P=0.148),然后进行参数估计(BIC=-4.293)和白噪声检验(Q=22.150,P=0.138),据此建立ARIMA模型,ARIMA(0,0,1)(0,1,1)12模型为最优模型,能较好的模拟甲型病毒性肝炎的发病。结论 ARIMA(0,0,1)(0,1,1)12模型能较好的模拟甲肝发病在时间序列的变化趋势,为制定科学的防控措施和策略提供依据。

Objective To explore the application of time series model to predicting viral hepatitis A （ VHA ） so as to provide a scientific basis for adopting the prevention and control measures in the next step. Methods According to the monthly incidence of VHA in Yichang City from 2005 to 2015, 2 kinds of models were built to predict the incidence of VHA in 2016. The agreement between the predicted value and the actual value of VHA incidence was evaluated. Results ARIMA model required the stable data. The incidence of VHA in Yichang City showed seasonal fluctuations and belonged to unstable data firstly, but the data about the years of 2010-2015 were stable. The monthly incidence of VHA in Yiehang City during 2010-2015 was treated by seasonal difference （Z= 1.447,P= 0.148） , then the parameters were estimated（ BIC =-4.293） and white noise test was made （Q = 22.150, P = 0.138） , after that the ARIMA model was built. The optimal model was the ARIMA （ 0, 0, 1 ） （ 0, 1, 1 ） 12 model, which could well simulate incidence of VHA. Conclusions The changing trend of time series of VHA incidence can be simulated well using the ARIMA （ 0, 0, 1 ） （ 0, 1, 1 ） 12 model, and it can provide a basis for formulating scientific prevention and control measures as well as strategies.