黄晓霞, 张顺先, 赵俊伟, 司拨云, 王世文, 王英. 采用自回归移动平均模型预测中国手足口病月发病率[J]. 疾病监测, 2013, 28(5): 396-399. DOI: 10.3784/j.issn.1003-9961.2013.5.017
引用本文: 黄晓霞, 张顺先, 赵俊伟, 司拨云, 王世文, 王英. 采用自回归移动平均模型预测中国手足口病月发病率[J]. 疾病监测, 2013, 28(5): 396-399. DOI: 10.3784/j.issn.1003-9961.2013.5.017
HUANG Xiao-xia, ZHANG Shun-xian, ZHAO Jun-wei, SI Bo-yun, WANG Shi-wen, WANG Ying. Prediction of monthly hand foot and mouth disease incidence in China by using autoregressive integrated moving average model[J]. Disease Surveillance, 2013, 28(5): 396-399. DOI: 10.3784/j.issn.1003-9961.2013.5.017
Citation: HUANG Xiao-xia, ZHANG Shun-xian, ZHAO Jun-wei, SI Bo-yun, WANG Shi-wen, WANG Ying. Prediction of monthly hand foot and mouth disease incidence in China by using autoregressive integrated moving average model[J]. Disease Surveillance, 2013, 28(5): 396-399. DOI: 10.3784/j.issn.1003-9961.2013.5.017

采用自回归移动平均模型预测中国手足口病月发病率

Prediction of monthly hand foot and mouth disease incidence in China by using autoregressive integrated moving average model

  • 摘要: 目的 采用自回归移动平均(autoregressive integrated moving average, ARIMA)模型对中国(未含香港、澳门和台湾地区)的手足口病月发病率进行预测,为手足口病预防控制提供参考依据,为ARIMA在传染病预防控制中的运用提供新的领域。 方法 根据2008-2011年全国手足口病月报告发病率时间序列,以2012年1-7月的月发病率作为验证数据,建立中国手足口病月发病率的ARIMA模型。 结果 我国手足口病月发病率模型为ARIMA(1,0,0)(0,1,0)12,模型自回归参数AR1=0.779 (t=7.315,PQ=10.328,P=0.889),提示残差为白噪声。2012年1-7月实际值与预测值的相对误差平均值为28.62%,最大44.57%,最小4.92%。 结论 ARIMA可用于我国手足口病月发病率的预测,模型预测效果的优化有待原始数据进一步积累。

     

    Abstract: Objective To predict the monthly incidence of hand foot and mouth disease (HFMD) in the mainland of China by using autoregressive integrated moving average (ARIMA) model and provide evidence for the prevention and control of HFMD and more application of ARIMA model in communicable disease prevention and control. Methods According to the time series of reported monthly incidence of HFMD in China from 2008 to 2011, the ARIMA model predicting monthly HFMD incidence in China was established with the incidence of HFMD from January to July 2012 as demonstration data. Results The model predicting monthly incidence of HFMD in China is ARIMA (1,0,0,) (0,1,0)12, in which autoregressive (AR1) is 0.779 (t=7.315,PQ has no statistical significance (Ljung-Box Q=10.328, P=0.889) and residuals is the white noise. The average of the relative error between actual and predicted values from January to July in 2012 is 28.62% (maximum 44.57%, minimum 4.92%). Conclusion ARIMA can be used in the prediction of monthly HFMD incidence. More original data are needed in order to optimize the model.

     

/

返回文章
返回