SARIMA模型在大连市梅毒发病预测中的应用  

Application of SARIMA model in predicting syphilis incidence in Dalian City

作　　者：孙巍[1] 白金剑 安庆玉[1] 熊帮洁[1] 郭俐男 吴隽 SUN Wei;BAI JinJian;AN Qingyu;XIONG Bangjie;GUO Linan(Dalian Center for Disease Control and Prevention,Dalian,Liaoning 116000,China)

机构地区：[1]大连市疾病预防控制中心,辽宁大连116035 [2]不详

出　　处：《公共卫生与预防医学》2025年第3期60-63,共4页Journal of Public Health and Preventive Medicine

摘　　要：目的 探讨季节性差分自回归滑动平均模型(SARIMA)在大连市梅毒发病预测中的应用,为该市梅毒防控提供参考依据。方法 应用SPSS 25.0软件对大连市2010—2019年梅毒月报告监测数建立SARIMA模型,用该模型对2020年1~12月梅毒发病数进行预测,并与实际报告数进行比较,以评价模型预测结果。结果 SARIMA(2,1,0)×(0,1,1)_(12)为最优预测模型(P<0.001),统计量遵循贝叶斯信息准则(BIC)值最小原则,白噪声检验Liung-Box统计量差异均无统计学意义,残差为白噪声序列,以此模型预测的2020年各月发病数与实际发病数比较,平均绝对百分比例误差MAPE=8.93%,模型预测效果较好。结论 SARIMA(2,1,0)×(0,1,1)_(12)模型能较好拟合大连市梅毒发病情况,可用于大连市梅毒发病数的预测。Objective To investigate the application of seasonal analysis time series autoregressive moving average model(SARIMA) in predicting syphilis incidence in Dalian,and to provide reference for the prevention and control of syphilis.Methods SPSS25.0 software was used to establish a SARIMA model based on the monthly monitoring data of syphilis reported in Dalian from 2010 to 2019.The model was used to predict the number of syphilis cases in January-December 2020,and the results were compared with the actual reports to evaluate the prediction effectiveness.Results SARIMA(2,1,0)×(0,1,1)_(12) was the best prediction model(P=0.018),and the statistics followed Akaike Criterion(AIC) and Bayesian(BIC) principle of minimum value.There were no statistically significant differences between white noise test and Liung-Box statistics,and the residual was white noise sequence.Compared with the actual incidence in 2020,the mean absolute percentage error(MAPE) of the predicted incidence in each month by this model was 8.93%,indicating that the model had a good predictive effect.Conclusion SARIMA(2,1,0)×(0,1,1)_(12) model can well fit the incidence of syphilis in Dalian and can be used to predict the number of syphilis cases in Dalian.

关 键 词：梅毒 季节性差分自回归滑动平均模型 预测 

分 类 号：R181[医药卫生—流行病学]