论文部分内容阅读
目的评价ARIMA和GM(1,1)模型在细菌性痢疾发病预测的应用效果,为选择适宜的预测方法提供依据。方法利用深圳沙井街道2006-2013年的细菌性痢疾的发病数据构建ARIMA模型和GM(1,1)模型,评价拟合效果。结果建立的ARIMA(1,1,1)(0,1,1)12模型为:(1-0.623B)(1-B)(1-B12)Yt=(1-0.963B)(1-0.72B12)et,GM(1,1)模型为:Y(t+1)=-190.506e-0.23003t+250.126。两个模型的平均误差率(MER)分别为2.91%和6.24%;决定系数分别为0.994和0.967。结论 ARIMA模型对细菌性痢疾发病率预测的效果较好,但在预测传染病发病的时候,尽可能地使用多种模型进行拟合,挑选出效果最好的。
Objective To evaluate the application effects of ARIMA and GM (1,1) models in predicting the incidence of bacterial dysentery and provide the basis for selecting suitable prediction methods. Methods ARIMA model and GM (1,1) model were constructed based on the data of bacterial dysentery in Shenzhen Shajing Street from 2006 to 2013 to evaluate the fitting effect. Results The ARIMA (1,1,1) (0,1,1) 12 model was established as follows: (1-0.623B) (1-B) (1-B12) Yt = (1-0.963B) (1-0.72 B12) et, GM (1,1) model is: Y (t + 1) = - 190.506e-0.23003t + 250.126. The average error rates (MER) of the two models were 2.91% and 6.24%, respectively; the coefficients of determination were 0.994 and 0.967, respectively. Conclusions The ARIMA model is effective in predicting the incidence of bacillary dysentery. However, when predicting the onset of infectious diseases, the models should be used as much as possible and the best one should be selected.