摘要
目的将ARIMA模型运用于湖北省的肺结核病发病预测,为湖北省结核病防治与预警系统提供决策依据。方法运用SPSS 22. 0统计软件对湖北省2001年到2016年肺结核的数据进行了基本趋势分析后建起了ARIMA时间序列模型,并对模型进行了检验,预测出了2017、2018年的肺结核发病数。结果模型ARIMA(1,1,2)(1,1,0)12所有参数均通过统计学检验;Box-Ljung检验统计量Q=20. 156,P> 0. 05,残差序列为白噪声;拟合优度指标平稳R2为0. 287,贝叶斯信息准则(BIC)为12. 114,模型拟合精度较好。结论在对湖北省肺结核病发病情况的近期预测中,引入时间序列模型的ARIMA模型分析方法,能够为结核病防制工作提供科学依据。
Objective To describe the method and procedure of fitting time series with ARIMA model, and apply the model to the prediction of tuberculosis incidence in Hubei Province, so as to provide decision-making evidence for tuberculosis prevention and early warning system in Hubei Province. Methods Using SPSS 22.0 statistical software to analyze the basic trend of pulmonary tuberculosis data fi＇om 2001 to 2016 in Hubei Province, ARIMA time series model was established and tested, and the incidence of pulmonatry tuberculosis in 2017 and 2018 was predicted. Results All parameters of ARIMA （ 1,1,2） （ 1,1,0） ~2 passed statistical test ; Box-Ljung test statistic Q = 20. 156, P 〉 0. 05, residual sequence was white noise ; goodness-of-fit index stationary- R2 was 0. 287, and Bayesian information criterion （BIC） was 12. 114. The fitting accuracy of the model was good. Conclusion In the short-term prediction of the incidence of pulmonary tuberculosis in Hubei Province, the introduction of time series model of ARIMA model analysis method, can play- a more positive guiding significance.
作者
李家琦
王雷
宋媛媛
熊甜
胡樱
LI Jiaqi;WANG Lei;SONG Yuanyuan;XIONG Tian;HU Ying(School of Health Sciences,Wuhan University,Wuhan 430071,China;Hubei Provincial Center for Disease Control and Prevention,Wuhan 430079,China;Center of Global Health,Wuhan University,Wuhan 430071,China)
出处
《公共卫生与预防医学》
2018年第5期37-40,共4页
Journal of Public Health and Preventive Medicine
基金
中华预防医学会疫苗与免疫青年人才托举项目(Q2017A4201)
