摘要
收集2004—2019年北京市猩红热月发病人数和人口学资料,采用描述性统计方法和Joinpoint回归调查猩红热的流行病学变化趋势。北京市猩红热发病的平均年度百分比变化为(AAPC=1.866,95%CI:-2.968~6.941;t=0.816,P=0.428),流行趋势总体保持稳定,每年的4—6月和11—12月为发病高峰,呈双季节模式。最优SARIMA模型和最优ETS模型预测的平均绝对误差(MAD)、均方根误差(RMSE)、平均绝对百分比误差(MAPE)、平均误差率(MER)、方根百分比误差(RMSPE)五个误差指标分别为0.586、0.623、0.751、0.296、0.785和0.318、0.282、0.438、0.282、0.338,可见ETS模型的预测准确性高于SARIMA模型,可用来对北京市猩红热流行趋势进行预测预警,从而为猩红热动态精准化防控提供参考依据。
Collect monthly incidence and demographic data of scarlet fever in Beijing from 2004 to 2019,and use descriptive statistical methods and Joinpoint regression to investigate the epidemiological trends of scarlet fever.The average annual percentage change of scarlet fever incidence in Beijing is(AAPC=1.866,95%CI:-2.968~6.941;t=0.816,P=0.428),and the overall trend of the epidemic remains stable.The peak of incidence is from April to June and November to December each year,showing a bi seasonal pattern.The average absolute error(MAD),root mean square error(RMSE),mean absolute percentage error(MAPE),mean error rate(MER),and root mean percentage error(RMSPE)predicted by the optimal SARIMA and ETS models are 0.586,0.623,0.751,0.296,0.785,and 0.318,0.282,0.438,0.282,and 0.338,respectively.It can be seen that the prediction accuracy of the ETS model is higher than that of the SARIMA model,It can be used to predict and warn the trend of scarlet fever in Beijing,providing a reference basis for the dynamic and precise prevention and control of scarlet fever.
作者
柴峰
Chai Feng(Tiantan Community Health Service Center in Dongcheng District,Beijing 100062)
出处
《科技与健康》
2024年第5期125-128,共4页
Technology and Health
关键词
基于状态空间的指数平滑模型
季节性差分自回归滑动平均模型
猩红热
预测
性能比较
exponential smoothing model based on state space
seasonal difference autoregressive moving average model
scarlet fever
prediction
performance comparison