中华预防医学杂志
中華預防醫學雜誌
중화예방의학잡지
CHINESE JOURNAL OF
2011年
12期
1108-1111
,共4页
朱猛%祖荣强%霍翔%鲍昌俊%赵杨%彭志行%喻荣彬%沈洪兵%陈峰
硃猛%祖榮彊%霍翔%鮑昌俊%趙楊%彭誌行%喻榮彬%瀋洪兵%陳峰
주맹%조영강%곽상%포창준%조양%팽지행%유영빈%침홍병%진봉
模型,统计学%流感,人%预测%时间序列分析%自回归求积移动平均模型
模型,統計學%流感,人%預測%時間序列分析%自迴歸求積移動平均模型
모형,통계학%류감,인%예측%시간서렬분석%자회귀구적이동평균모형
Models,statistical%Influenza,human%Forecasting%Time series analysis%Autoregressive integrated moving average model
目的 探讨时间序列分析的自回归求积移动平均模型(ARIMA)在江苏省流行性感冒(流感)疫情预测预警中的应用.方法 收集江苏省2005年10月至2010年2月各监测点流感样病例(ILI)监测数据,建立基础数据库,对每周ILI发病人数进行ARIMA建模拟合,对2010年3-4月各周流感样病例发病情况进行预测.结果 构建得到ILI发病ARIMA(2,1,2)预测模型为(1+0.785B2)(1 -B)In Xt=(1+0.622B2)εt,其中,B代表后移算子,Xt代表流感样病例发病人数,εt为随机误差.模型残差序列为白噪声,模型的Ljung-Box检验统计量为5.087,P=0.995,拟合效果良好.应用该模型预测2010年3-4月江苏省各周ILI发病情况,预测值符合实际发病变动趋势,实际观测值均在预测值95％CI值内.结论 ARIMA模型能较好地模拟江苏省ILI发病情况.
目的 探討時間序列分析的自迴歸求積移動平均模型(ARIMA)在江囌省流行性感冒(流感)疫情預測預警中的應用.方法 收集江囌省2005年10月至2010年2月各鑑測點流感樣病例(ILI)鑑測數據,建立基礎數據庫,對每週ILI髮病人數進行ARIMA建模擬閤,對2010年3-4月各週流感樣病例髮病情況進行預測.結果 構建得到ILI髮病ARIMA(2,1,2)預測模型為(1+0.785B2)(1 -B)In Xt=(1+0.622B2)εt,其中,B代錶後移算子,Xt代錶流感樣病例髮病人數,εt為隨機誤差.模型殘差序列為白譟聲,模型的Ljung-Box檢驗統計量為5.087,P=0.995,擬閤效果良好.應用該模型預測2010年3-4月江囌省各週ILI髮病情況,預測值符閤實際髮病變動趨勢,實際觀測值均在預測值95％CI值內.結論 ARIMA模型能較好地模擬江囌省ILI髮病情況.
목적 탐토시간서렬분석적자회귀구적이동평균모형(ARIMA)재강소성류행성감모(류감)역정예측예경중적응용.방법 수집강소성2005년10월지2010년2월각감측점류감양병례(ILI)감측수거,건립기출수거고,대매주ILI발병인수진행ARIMA건모의합,대2010년3-4월각주류감양병례발병정황진행예측.결과 구건득도ILI발병ARIMA(2,1,2)예측모형위(1+0.785B2)(1 -B)In Xt=(1+0.622B2)εt,기중,B대표후이산자,Xt대표류감양병례발병인수,εt위수궤오차.모형잔차서렬위백조성,모형적Ljung-Box검험통계량위5.087,P=0.995,의합효과량호.응용해모형예측2010년3-4월강소성각주ILI발병정황,예측치부합실제발병변동추세,실제관측치균재예측치95％CI치내.결론 ARIMA모형능교호지모의강소성ILI발병정황.
Objective This research aimed to explore the application of ARIMA model of time series analysis in predicting influenza incidence and early warning in Jiangsu province and to provide scientific evidence for the prevention and control of influenza epidemic.Methods The database was created based on the data collected from monitoring sites in Jiangsu province from October 2005 to February 2010.The ARIMA model was constructed based on the number of weekly influenza-like illness (ILI) cases.Then the achieved ARIMA model was used to predict the number of influenza-like illness cases of March and April in 2010.Results The ARIMA model of the influenza-like illness cases was ( 1 +0.785B2) (1 -B) In Xt =(1 + 0.622B2) εt.Here B stands for back shift operator,t stands for time,Xt stands for the number of weekly ILI cases and εt stands for random error.The residual error with 16 lags was white noise and the Ljung-Box test statistic for the model was 5.087,giving a P-value of 0.995.The model fitted the data well.True values of influenza-like illness cases from March 2010 to April 2010 were within 95％ CI of predicted values obtained from present model.Conclusion The ARIMA model fits the trend of influenza-like illness in Jiangsu province.
