ARIMA模型在河南省甲型病毒性肝炎发病数预测中的应用

发布时间：2018-06-30 08:41

本文选题：乘积季节自回归移动平均模型 + 甲型病毒性肝炎　；参考：《现代预防医学》2017年07期

【摘要】：目的建立乘积季节自回归移动平均(ARIMA)模型,利用该模型预测河南省甲肝发病情况并探讨其可行性。方法对2008年1月-2015年8月河南省的甲肝疫情监测数据差分平稳化,通过专家建模器筛选最优模型,利用2015年9月-2016年8月的甲肝疫情资料来拟合预期值并评价该模型的可行性。结果 2008-2015年河南省甲肝发病数逐年减少且呈现明显的季节效应;本次研究中乘积季节ARIMA(0,1,1)×(1,0,1)_(12)模型能较好的拟合既往的甲肝报告病例数,模型统计量Ljung-Box Q为21.742,P为0.115〉0.05,残差序列为白噪音;且对2015年9月-2016年8月按月报告的甲肝病例数的预测值与实际值吻合情况良好,平均误差绝对值4.67,平均相对误差绝对值为0.2。结论ARIMA模型能较好模拟、预测河南省甲肝的发病情况,该模型的预测效能将优化甲肝预防工作,有较好的推广价值。
[Abstract]:Objective to establish a seasonal autoregressive moving average (Arima) model for predicting the incidence and feasibility of hepatitis A in Henan province. Methods the data of hepatitis A epidemic surveillance in Henan Province from January 2008 to August 2015 were compared and stable. The optimal model was screened by expert modeler, and the expected value was fitted with the data from September 2015 to August 2016, and the feasibility of the model was evaluated. Results from 2008 to 2015, the incidence of hepatitis A in Henan Province decreased year by year and showed obvious seasonal effect. The model statistic Ljung-Box Q is 21.742 (P = 0.115 > 0.05), the residual sequence is white noise, and the predicted value of the monthly reported cases of hepatitis A from September 2015 to August 2016 is in good agreement with the actual value. The average absolute error is 4.67 and the average relative error absolute value is 0.2. Conclusion Arima model can be used to predict the incidence of hepatitis A in Henan province. The prediction efficiency of Arima model will optimize the prevention of hepatitis A and have a good value of popularization.
【作者单位】：郑州大学公共卫生学院;河南省疾病预防控制中心;
【基金】：郑州大学青年教师启动基金(32210273)
【分类号】：R512.61

【相似文献】