山西省痢疾发病趋势的时间序列分析

目的探讨季节性自回归移动平均混合(SARIMA)模型分析预测山西省痢疾发病率的可行性和适用性,为痢疾的预防与控制提供决策依据。方法利用R 3.3.1对山西省2004年1月—2013年12月痢疾月发病率资料进行建模,并以2014年痢疾月发病率资料验证模型的预测效果。结果模型较好地拟合了山西省痢疾月发病率,模型残差为白噪声序列,预测值与实际值的相对误差范围为0.909%~35.575%,平均相对误差为13.399%。结论 SARIMA模型可较好地反映山西省痢疾的发病趋势并进行短期预测。

Time series analysis on incidence trend of dysentery in Shanxi Province

Objective To investigate the feasibility and applicability of model of seasonal autoregressive integrated moving average(SARIMA)in analysis and prediction on dysentery morbidity in Shanxi Prov-ince,and to provide decision basis for preventing and controlling dysentery.Methods SARIMA model was established based on monthly morbidity of dysentery from Jan.2004 to Dec.2013 by using R3.3.1 software,and the predictive effect of the model was verified with monthly morbidity of dysentery in 2014. Results SARIMA model(2,1,2)(1,1,1)12fitted very well with the monthly dysentery morbidity in Shanxi Province,and the residuals of the model were white noise sequence.The relative error between predicted and actual values at a range from 0.909% to 35.575% and average relative error was 13.399%.Conclusion SARIMA model can well reflect the incidence trend of dysentery and predict the morbidity of dysentery in a short term in Shanxi Province.