深圳市登革热传播风险ARIMA模型预测分析

ARIMA model analyzing and predicting to epidemic tendency of dengue fever in Shenzhen

目的利用拟合ARIMA模型对深圳市登革热发病趋势进行时间序列分析和预测,为制定登革热防治策略提供科学依据。方法收集深圳市2011-2015年登革热月发病资料,通过SPSS 20.0和SAS 9.4统计软件拟合ARIMA模型,预测2015年7-12月的发病率。结果最终拟合为ARIMA（1,0,0）（0,1,0）12模型,残差为白噪声序列,预测值与实际值的平均相对误差为20.6%。2015年7、8、9月份登革热发病的预测值符合实际值的变动趋势。结论 ARIMA模型能较好模拟深圳市登革热的短期发病趋势。

Objective To analyze and predict incidence of dengue fever（DF） in Shenzhen with an auto regressive integrated moving average（ARIMA） model,which may provide a scientific basis for developing prevention and control strategies on DF in the future. Methods Incidence of DF was collected monthly from 2011 to 2015 and a model（ARIMA） was fit with SPSS 20.0 and SAS9.4 software, to predict the incidence of DF from July to December 2015. Results The model ARIMA（1, 0, 0）（ 0, 1, 0）12 was established finally and the residual sequence was a white noise sequence. The relative error in average was 20.6%between the forecasting value and the real value. The predicted incidence in July to September 2015 was consistent with the actual one. Conclusion ARIMA model can predict the shot-time trend of DF in Shenzhen.