应用3种回归模型预测手足口病周发病数

发布时间：2018-05-11 17:50

  本文选题：自回归 + 季节性　； 参考：《现代预防医学》2016年16期

【摘要】：目的探索分析手足口病周数据的统计学方法,提升手足口病预测能力。方法中国疾病预防控制信息系统导出2008年第1周至2014年第14周北京市通州区手足口病周发病数。采用SPSS 17.0软件进行自回归、季节性自回归与混合Serfling回归模型拟合。结果自回归、季节性自回归、混合Serfling回归3种模型对2008年第1周至2014年第14周实际发病数进行拟合,回归方程R2分别是0.907、0.917、0.919,所得残差经Ljung-Box检验均是白噪声;以所得回归方程对2014年第15周至第38周实际发病数进行预测,3种模型的平均绝对百分比误差(MAPE)分别为:18.67%、18.43%、17.12%。结论混合Serfling回归模型预测效果最优。
[Abstract]:Objective to explore the statistical method of analyzing the week data of hand, foot and mouth disease (HFMD) in order to improve the predictive ability of HFMD. Methods from the first week of 2008 to the 14th week of 2014, the incidence of hand, foot and mouth disease in Tongzhou District of Beijing was derived from China Disease Prevention and Control Information system. The software SPSS 17.0 was used to carry out autoregressive, seasonal autoregressive and mixed Serfling regression models. Results autoregressive, seasonal autoregressive and mixed Serfling regression models were used to fit the actual incidence from the first week of 2008 to the 14th week of 2014. The regression equation R2 was 0.9070.917 ~ 0.919, respectively. The residual errors were all white noise by Ljung-Box test. According to the regression equation, the average absolute percentage error (MAPE) of the three models for predicting the actual incidence from week 15 to week 38 of 2014 is: 1: 18.67 and 18.43 and 17.12 respectively. Conclusion the mixed Serfling regression model has the best prediction effect.
【作者单位】： 北京市通州区疾病预防控制中心;
【分类号】：R725.1;R181.3
，

本文编号：1874955