基于ARIMA 模型的贵州省2017年流行性乙型脑炎发病预测

目的 探讨ARIMA模型用于预测流行性乙型脑炎发病的可行性，并利用模型预测贵州省2017年乙脑发病趋势。方法 采用SPSS20.0对贵州省2007 - 2016年乙脑报告病例数进行分析并构建ARIMA模型，使用筛选的最优模型预测贵州省2017年乙脑发病。结果 ARIMA（0,0,1）（1,1,0）12模型为最优模型，模型BIC = 6.769，稳定的R2 = 0.496；除常数项外，该模型各参数均有统计学意义，残差序列为白噪声序列。用该模型拟合贵州省2008 - 2016年乙脑月发病数，拟合数与报告数的变化趋势基本相同，差异均无统计学意义（Ｐ＞0.05）；贵州省2017年乙脑预测病例数为41例，高峰仍在7、8月。结论 ARIMA模型可以较好地拟合贵州省乙脑的发病趋势，可用于短期预测；与2016年相比，预测2017年乙脑发病相对平稳。

The prediction of the Japanese encephalitis invasion based on the ARIMA model in Guizhou in 2017

Objective The aim of this study was to explore the application of ARIMA model in prediction of the Japanese encephalitis invasion, and predict the invasion trend of Japanese encephalitis in Guizhou in 2017. Methods The reported data of Japanese encephalitis invasion from 2007 to 2016 in Guizhou province was analyzed, and an ARIMA model was established by SPSS20.0 statistical software. The possible numbers of Japanese encephalitis in 2017 were predicted by the best model screened. Results ARIMA（0,0,1）（1,1,0）12 model was the best one, of which BIC was 6.769 and R2 was 0.496. Except the constant term, there was a statistical significance in the parameters of the model. The model residual sequence was white noise. The variation tendency of the fitted value by the model from 2008 to 2016 was in accordance with the reported cases, and the difference between the fitted value and actual value had no statistical significance (Ｐ>0.05). 41 cases of the Japanese encephalitis were predicted in 2017, and the peak would be still at July and August. Conclusion ARIMA model can be used to fit the incidence trends of Japanese encephalitis in Guizhou province and obtaun a short-term prediction. The incidence of Japanese encephalitis would be relatively stable in 2017 compared to 2016.