【摘要】 目的 了解2010—2024年玉林市乙型病毒性肝炎（乙肝）分布情况，构建ARIMA模型和SVR模型预测乙肝发病趋势，为制订玉林市乙肝防控措施提供科学依据。方法 采用描述性流行病学方法对2010—2024年玉林市乙肝发病情况进行分析，使用R 4.5.0软件构建ARIMA模型和SVR模型预测乙肝发病趋势。结果 2010—2024年玉林市乙肝共报告129 959例，年均报告发病率为151.472/10万，整体呈下降趋势（趋势χ²=5 093.172,P<0.001）。高发地区为兴业县、博白县和福绵区（χ²=2 337.990,P<0.001）;1月报告病例数最多（Z=73 371.933,P<0.001）；男、女性别比为2.256∶1，年龄集中在20～<60岁（占82.24%），发病人群以农民为主（80.67%）。使用ARIMA模型和SVR模型预测2024年1—12月的MAE值分别为199.664、281.640,RMSE值分别为213.198、339.453,MAPE值分别为63.020%、101.653%。结论 玉林市乙肝发病率呈逐年下降趋势，男性、20～<60岁人群、农民为乙肝防控的重点人群，使用ARIMA（0,1,1）（0,1,2）₁₂模型更能精准预测未来一段时期内乙肝的发病数。更多还原

【Abstract】 Objective To understand the distribution of viral hepatitis B(hereinafter referred to as hepatitis B) from 2010 to 2024 in Yulin, and to construct the ARIMA and SVR models for predicting the incidence trend of hepatitis B, providing a scientific basis for hepatitis B control and prevention strategies in Yulin. Methods Descriptive epidemiological method was adopted to analyze incidence of hepatitis B from 2010 to 2024 in Yulin. R 4.5.0 software was employed to construct the ARIMA and SVR models for predicting incidence trend of hepatitis B. Results A total of 129 959 hepatitis B cases were reported in Yulin from 2010 to 2024, with an average annual reported incidence rate of 151.472/100 000, showing an overall downward trend(trend χ~2=5 093.172, P<0.001). The areas with high incidences were Xingye County, Bobai County and Fumian District(χ~2=2 337.990, P<0.001). The number of reported cases was the highest in January(Z=73 371.933, P<0.001), the male to female ratio was 2.256∶1, the age of patients was mainly 20-<60 years old(82.24%), and farmers were the main incidence population of the cases(80.67%). Using the ARIMA and SVR models to predict the MAE of January to December 2024 were 199.664 and 281.640, RMSE were 213.198 and 339.453, and MAPE were 63.020% and 101.653%, respectively. Conclusions The incidence rate of hepatitis B in Yulin indicates a decreasing trend year by year. Males, 20-<60 years old group and farmers are the key groups for hepatitis B control and prevention. Employing ARIMA(0,1,1)(0,1,2) 12 model can more accurately predict the incidence number of hepatitis B in the coming period.更多还原