南通市流行性腮腺炎疫情的预测预警分析

发布时间：2018-03-24 14:32

本文选题：时间序列分析　切入点：ARIMA模型　出处：《南京医科大学学报(自然科学版)》2014年11期

【摘要】：目的 :探讨时间序列分析中的求和自回归移动平均模型(autoreg ressive integrated moving average,ARIMA)乘积季节模型在南通市流行性腮腺炎疫情预测预警中的应用。方法:收集"中国疾病预防控制系统"中2004年1月~2013年11月流行性腮腺炎的月发病监测数据,建立时间序列数据库,对每月腮腺炎的发病人数进行ARIMA模型拟合,利用模型对2004年1月~2013年11月的数据进行回代预测,并对2013年12月~2014年6月各月的流行性腮腺炎的发病情况进行前瞻性预测。结果 :构建流行性腮腺炎的ARIMA(0,1,2)(1,1,0)12模型为(1-B)(1+0.51B12)caset=(1-0.24B-0.29B2)(1-B12)εt,其中B代表后移算子,caset代表年腮腺炎发病数,εt为随机误差。回代预测的实际值和预测值基本相符,均在95%可信区间内。前瞻性预测结果符合流行性腮腺炎的流行特征。结论:该模型能较好模拟并预测腮腺炎的发病情况,流行性腮腺炎发病数有增加趋势,应进一步分析本地腮腺炎的流行特征及其免疫策略。
[Abstract]:Objective: to explore the application of autoreg ressive integrated moving moving Arima model in time series analysis to the prediction and early warning of epidemic situation of mumps in Nantong City. Monthly surveillance data of mumps from January 2004 to November 2013, Time series database was established to fit the incidence of mumps by ARIMA model. The data from January 2004 to November 2013 were forecasted by the model. The incidence of mumps in the months from December 2013 to June 2014 was forecasted. Results: the model of mumps was established as 1. 51B12 casettir 1-0.24B-0.29B2 (1-B12) 蔚 t, where B represents the backward shift operator caset for the incidence of annual parotitis. , 蔚 t is random error. The actual value of the back generation prediction is basically consistent with the predicted value. The predicted results were consistent with the epidemic characteristics of mumps. Conclusion: this model can well simulate and predict the incidence of mumps, and the incidence of mumps tends to increase. The epidemic characteristics and immunization strategy of local mumps should be further analyzed.
【作者单位】：南通市疾病预防控制中心;
【分类号】：R181.8;R512.1

【参考文献】