基于ARIMA的传染病流行趋势预测及防治对策

基于ARIMA的传染病流行趋势预测及防治对策

桂腾叶1,陈硕2,隗立志2,崔志军2,周晓琳2

(1.西藏大学 工学院,西藏 拉萨850000;2.西藏大学 藏文信息技术研究中心,西藏 拉萨850000)

摘要为准确地预测传染病,根据传染病变化特点,提出了一种差分自回归移动平均模型的传染病预测模型。模型对原始数据进行平稳化预处理,消除其突发性、季节性和周期性特征,并利用ARIMA对将平稳后的数据进行建模,采用某市流行性感冒发病率数据进行仿真,实验结果表明,ARIMA模型能较好地捕捉传染病变化规律,提高了预测精度,是一种有效预测传染病的方法,同时能为传染病的预防监测措施提供决策依据。

关键词传染病;平稳时间序列;预处理;ARIMA

收稿日期:2015-04-28

基金项目:西藏大学自治区大学生创新性实验训练计划基金资助项目(2014QCX049)

作者简介:桂腾叶(1991—),男,本科。研究方向:交通运输规划与管理。E-mail:1148195268@qq.com

doi:10.16180/j.cnki.issn1007-7820.2015.12.013

中图分类号O212;TP274文献标识码A

Prediction and Countermeasures of Infectious Diseases Based on ARIMA

GUI Tengye1,CHEN Shuo2,WEI Lizhi2,CUI Zhijun2,ZHOU Xiaolin2

(1.School of Engineering,University of Tibet,Lhasa 850000,China;

2.Tibetan Information Technology Research Center,University of Tibet,Lhasa 850000,China)

AbstractAn autoregressive integrated moving average model (ARIMA) based on the characteristics of infectious diseases’ changes is proposed for accurate prediction of the epidemics.Firstly,the original data is preprocessed smoothly by the model,which is also used to eliminate the abruptness and the characteristics of seasonal and cyclical.Then the ARIMA is used to build a model of data that has been smoothed.We use the data,the morbidity of influenza in a city,to simulate the results of the experiment,which shows that ARIMA can capture the regulation of the variation of infectious diseases well,and improve the accuracy of the prediction.It is an efficient approach to predicting the epidemic,which also provides support for decision-making in epidemic prediction and monitoring.

Keywordsinfectious disease;stationary time series;pretreatment;ARIMA

当前,对传染病有许多预测方法和模型[1]。而各种方法具有不同优缺点。传染病由于受到各种环境因素和自身的身体条件的影响,想要对传染病进行准确的预测,就必须同时考虑这些因素。综合多方面情况,笔者选择差分自回归移动平均模型(ARIMA模型)进行预测[2-3],ARIMA模型在疾病控制[4]、经济预测[5]、交通量预测[6]和生态环境模拟与预测[7]等方面都有应用。是一种无需了解影响被预测变量的相关因素,可刻画具有时间变化特征,能够根据一些随时间变化而又相互关联的动态数据建立的一种效果较好的预测模型[8],从而为传染病预测提供了新的途径。……