采用自回归移动平均模型预测中国流感病例数

发布时间：2018-04-30 01:26

本文选题：流行性感冒(流感) + 月报告病例数　；参考：《病毒学报》2017年05期

【摘要】：采用自回归移动平均模型(ARIMA)对中国(不含中国港澳台地区)流感月报告病例数进行预测研究,为中国流行性感冒(流感)的预防控制提供参考依据。使用SPSS 24.0软件,以2006年1月至2016年12月中国流感月报告病例数建立时间序列模型,并以2017年1~5月的月报告病例数作为验证数据,评估和筛选最优模型。以2006年1月至2016年12月中国流感月报告病例数为基础数据,建立的最优模型为ARIMA(4,0,4),其平稳R2=0.672,标化BIC=18.032,Ljung-Box Q=16.381,P=0.089。对2017年1~5月的数据进行预测,预测相对误差的平均值仅为-3.25%。ARIMA模型在预测中国流感月报告病例数方面效果较好,但模型的建立和预测应用是个动态过程,需不断根据积累的数据进行调整,从而提高预测精度。
[Abstract]:An autoregressive moving average model (ARIMA) was used to predict the number of monthly influenza cases reported in China (excluding Hong Kong, Macao, Taiwan and China) in order to provide a reference for the prevention and control of influenza in China. Using SPSS 24.0 software, the time series model of monthly reported cases of influenza in China from January 2006 to December 2016 was established, and the monthly reported cases from January to May 2017 were used as validation data to evaluate and screen the optimal model. Based on the data of monthly reported cases of influenza in China from January 2006 to December 2016, the optimal model was established as Arima (4), 0.672 (stable R2), and (16.382) Ljung-Box (0.089), standardized BIC-18.032 (Ljung-Box). Based on the data from January to May 2017, the average relative error of prediction is only -3.25. Arima model is effective in predicting the number of monthly reported cases of influenza in China, but the establishment and application of the model is a dynamic process. It is necessary to adjust continuously according to the accumulated data so as to improve the accuracy of prediction.
【作者单位】：兰州大学第一医院老年呼吸科;河北省人民医院;中国疾病预防控制中心寄生虫病预防控制所;甘肃省疾病预防控制中心;中国疾病预防控制中心病毒病预防控制所;
【分类号】：R181.3;R511.7

【相似文献】