TBATS模型在流行性腮腺炎发病率预测中的应用

目的探讨TBATS模型在流行性腮腺炎发病率预测中的可行性。方法以2004—2017年江西省流腮数据作为演示数据。2004年1月至2017年6月数据作为训练集,2017年7~12月数据作为测试集。训练集分别训练TBATS模型、SARIMA模型,并预测2017年7~12月发病率并与测试集比较,采用MAPE,RMSE,MAE和MER评价模型拟合及预测效果。结果SARIMA(1,0,0)(1,1,0)₁₂含常数项为最优SARIMA模型。TBATS模型、SARIMA模型拟合的MAPE、MAE、RMSE和MER依次为15.06%,0.21,0.29,13.57%和21.93%,0.29,0.41,18.73%;TBATS模型、SARIMA模型预测的MAPE、MAE、RMSE和MER依次为7.95%,0.08,0.11,7.12%和15.33%,0.17,0.18,14.93%。结论TBATS模型对流行性腮腺炎发病率预测精度高,值得推广应用。

Application of TBATS in the prediction of mumps incidence

Objective To explore the applicability of the TBATS in predicting the incidence of mumps. Methods The incidence of mumps of Jiangxi Province from 2004 to 2017 was used as the demonstration data.The incidence of mumps in Jiangxi Province from July to December 2017 was used as test data.The training data from January 2004 to June 2017 were used to train the TBATS and the SARIMA,and predict the value from July to December 2017.The fitted and predicted values were compared with the test data.The MAPE,RMSE,MAE and MER were used to evaluate model fitting and prediction effects. Results SARIMA(1,0,0)(1,1,0)₁₂ with drift was the optimal SARIMA.The MAPE,MAE,RMSE and MER fitted by the TBATS and the SARIMA were 15.06%,0.21,0.29,13.57% and 21.93%,0.29,0.41,18.73%,respectively.The MAPE,MAE,RMSE and MER predicted by the TBATS and the SARIMA were 7.95%,0.08,0.11,7.12% and 15.33%,0.17,0.18,14.93%. Conclusion The TBATS has high accuracy in predicting the incidence of mumps and is worthy of popularization and application.