论文部分内容阅读
目的构建义乌市乙类传染病流行趋势最优预测模型。方法通过对义乌市2010-2012年乙类传染病按月发病率的日平均进行ARIMA时间序列、GM(1,1)灰色模型以及两者的组合模型进行建模,并预测2013年1-4月发病率与实际比较。结果 ARIMA(1,1,1)模型的r(曲)为82.49%,GM(1,1)灰色模型的r(曲)为59.07%,组合预测模型的r(曲)为70.89%;ARIMA(1,1,1)预测的MAPE为118.77%、MSPE为67.07%,GM(1,1)预测的MAPE为15.50%、MSPE为9.33%,组合预测模型预测的MAPE为107.24%、MSPE为60.06%。结论在进行义乌市乙类传染病发病流行趋势建模中,ARIMA是一个较好的方法,但是进行预测不如GM(1,1)。
Objective To construct the optimal prediction model of epidemic tendency of Group B infectious diseases in Yiwu City. Methods The ARIMA time series, GM (1,1) gray model and the combined model of the two models were modeled by the daily average monthly incidence of Group B infectious diseases in Yiwu from 2010 to 2012, Monthly incidence and actual comparison. Results The r (song) of the ARIMA (1,1,1) model was 82.49%, the r (song) of the GM (1,1) gray model was 59.07% and that of the combined forecasting model was 70.89% 1, 1, 1) predicted MAPE was 118.77%, MSPE was 67.07%, GM (1,1) predicted MAPE was 15.50%, MSPE was 9.33%, combined forecasting model predicted MAPE was 107.24%, MSPE was 60.06% . Conclusion ARIMA is a better method for modeling epidemic trend of infectious diseases in Yiwu City, but it is not as good as GM (1,1) in forecasting.