摘要
目的探讨求和自回归滑动平均模型(ARIMA)在全国肺结核月发病率中的应用,并利用该模型进行未来肺结核月发病率的预测。方法利用中国疾病预防控制中心公共卫生科学数据中心网站上公布的2004-2014年全国肺结核月发病率数据建立ARIMA模型,并预测2014年月发病率,并加入2014年的实际月发病率数据对模型进行修正,并预测2015年肺结核月发病率。结果 ARIMA(0,1,0)×(0,1,1)12模型能很好地拟合2004-2014年全国肺结核月发病率数据,模型的决定系数(R2)为0.984,预测2014年肺结核月发病率的效果尚可。加入2014年实际月发病率,修正后的最优模型为ARIMA(1,0,2)×(1,0,0)12,决定系数(R2)为0.935。结论 ARIMA模型能很好地对未来的肺结核月发病率进行预测,预测到2015年肺结核疫情基本处于平稳态势。
Objective To explore the application of auto regressive integrated moving average （ARIMA） model on the research of the monthly incidence of pulmonary tuberculosis, and to forecast the incidence trend of pulmonary tuberculosis. Methods ARIMA model was established based on the data of the monthly incidence of pulmonary tuberculosis from 2004 to 2014, and forecasted the monthly incidence of pulmonary tuberculosis in 2014. Then we refitted ARIMA model according to the monthly incidence of pulmonary tuberculosis from 2004 to 2014, and forecasted the monthly incidence of pulmonary tuberculosis in 2015. Results The ARIMA（0, 1, 0）×（0, 1, 1）12 model was established, and the determination coefficient （R2） was 0.984. The prediction effect of model was acceptable for monthly incidence of pulmonary tuberculosis in 2014. The re-fitted model was ARIMA （1, 0, 2）×（1, 0, 0）12 model, and the RE was 0.935. Conclusion The ARIMA model could well extract the information of the monthly incidence of pulmonary tuberculosis in the future, and the pulmonary tuberculosis epidemic situationi is basically stable in 2015.
出处
《中国公共卫生管理》
2016年第4期438-440,共3页
Chinese Journal of Public Health Management
基金
2015年皖南医学院中青年科研基金(WK201503)
关键词
肺结核
月发病率
ARIMA模型
时间趋势
pulmonary tuberculosis
incidence
auto regressive integrated moving average model
time trend
