时间序列ARIMA模型在乙型肝炎疫情预测中的应用

Forecasting Hapetitis B Epidemic Situation by Applying the ARIMA Model

目的应用时间序列模型中的自回归-求和-移动平均模型（Autoregressive Integrated Moving Average madel，ARIMA），分析乙肝发病，为乙肝的预警预测提供科学依据。方法利用上海市闸北区传染病监测系统7a多乙肝发病情况数据，建立乙肝疫情ARIMA预测模型。结果上海市闸北区2002～2009年乙肝时间序列符合ARIMA（1，1，1）模型（Ljung-Box检验，P=0．799），模型残差自相关系数均在±0．5之间，预测值与观测值具有较高的吻合度。结论乙肝ARIMA模型对乙肝疫情预测有较好的效果。

Objective To analyze the hapetitis B epidemic situation by applying the ARIMA model, and provide scientific evidence for forecasting hepatitis B epidemic. Methods The forecasting model was setup based on over seven years＇ data of the hepatitis B surveillence in shanghai zhabei district and in the application of ARIMA model. Results The hepatitis B time series of shanghai zhabei district accorded with ARIMA （1,1,1） model by Ljung-Box statistics （P=0.799）, and correlation coefficient of model residual was within ±0.5 The predicted value had better accord with the observed value. Conclusions It was practical to apply the approach of ARIMA model to predict hepatitis B epidemic.