Page 78 - 南京医科大学学报自然科学版

南京医科大学学报（自然科学版） 第44卷第1期
· 72 · Journal of Nanjing Medical University（Natural Sciences） 2024年1月

·流行病学研究·

我国2012—2021年4种肝炎流行趋势的时间序列分析和预测

马一鸣，丁 勇 *
南京医科大学康达学院医学信息工程学部，江苏 连云港 222000

［摘 要］ 目的：分析我国2012―2021年4种病毒性肝炎流行特征的季节性规律和长期趋势，探讨适合肝炎发病预测的时间
序列模型，为科学防控肝炎提供参考依据和建议。方法：对我国2012年1月—2021年12月甲型、乙型、丙型和戊型肝炎的月
发病例数进行时间序列的季节性分解，建立季节自回归移动平均模型（autoregressive integrated moving average model，ARIMA）
和季节指数平滑模型（exponential smoothing model，ES），并对2022年1―8月4种肝炎的发病例数进行预测，并比较预测效果。
结果：每年3月份是各类肝炎发病的高发期，10年期间，甲型肝炎总体保持下降趋势，乙型肝炎总体趋势有升有降，近年来有上
升趋势；丙型肝炎总体呈上升趋势；戊型肝炎总体保持平稳趋势。乙型、丙型和戊型肝炎月平均发病例数分别为甲型肝炎的
57.06倍、11.50倍、1.35倍。季节ES模型的预测效果要优于季节ARIMA模型。结论：我国乙型和丙型肝炎发病人数众多，要加
强重点防控。时间序列的季节性分解可用于分析肝炎流行特征的季节性规律和长期趋势，季节指数平滑模型中水平、趋势和
季节3个参数，能体现肝炎发病的流行规律，在肝炎发病预测中，具有模型简单、计算简便、预测精度高的优点。
［关键词］ 病毒性肝炎；时间序列；季节性分解；预测
［中图分类号］ R512.6 ［文献标志码］ A ［文章编号］ 1007⁃4368（2024）01⁃072⁃08
doi：10.7655/NYDXBNSN230896

Time series analysis and forecasting of four hepatitis epidemic trends in China from 2012 to
2021

MA Yiming，DING Yong *
Medical Information Engineering Department of Kangda College，Nanjing Medical University，Lianyungang
222000，China

［Abstract］ Objective：To analyze the seasonal patterns and long⁃term trends of the 10 year epidemic characteristics of four types of
viral hepatitis in China from 2012 to 2021，and explore a time series model suitable for forecasting predicting hepatitis incidence，
providing reference and suggestions for scientific hepatitis prevention and control. Methods：Seasonal decomposition of the time series
was conducted on the monthly incidence of hepatitis A，B，C，and E in China from January 2012 to December 2021. A seasonal
autoregressive integrated moving average model（ARIMA）and a seasonal index smoothing model（ES）were established to predict the
incidence of four types of hepatitis from January to August 2022，and the predictive effects were compared. Results：March of each year
is the peak period for the incidence of all types of hepatitis. Over the 10 year period，the hepatitis A showed an overall decreasing
trend，hepatitis B had fluctuating trends with recent years showing an increasing trend，hepatitis C showed an overall increasing trend，

and hepatitis E remained stable overall. The monthly average incidence of hepatitis B，C，and E were 57.06 times，11.5 times，and
1.35 times higher than that of hepatitis A，respectively. The prediction performance of the seasonal ES model was better than that of the
seasonal ARIMA model. Conclusion：There are a large number of patients with hepatitis B and C in China，and key prevention and
control efforts need to be strengthened. The seasonal decomposition of time series can be used to analyze the seasonal patterns and
long⁃term trends of hepatitis prevalance. The seasonal ES model includes three parameters：level，trend，and seasonality，which can
reflect the epidemic pattern of hepatitis. In the prediction of hepatitis incidence，it has the advantages of being simple，easy to
calculate，and high prediction accuracy.
［Key words］ viral hepatitis；time series；seasonal decomposition；forecast
［J Nanjing Med Univ，2024，44（01）：072⁃079］

［基金项目］ 南京医科大学康达学院第二期品牌专业建设工程资助项目（JX206000302）；南京医科大学康达学院医学信息模
拟及预测科研团队资助项目（KD2022KYCXTD003）
∗
通信作者（Corresponding author），E⁃mail：yding @ njmu.edu.cn