成都市人口死亡率时间序列分析与预测.pdfVIP

2 6 200 6 23 3 J PrevM ed Inf, Jun1200 , Vol123, No13 成都市人口死亡率时间序列分析与预测 , , [ ] , , , , Box- Ljung ARIMA , ARIAM ( 0, 1, 1) ( 0, 1, 1) , MAPE = 81 50% 2000, 12 2 , [ ] ; ; ARIMA [ ] R1951 1 [] A [] 1006- 4028( 200 ) 03- 2 6- 04 Time SeriesAnalysis and Forecast ofMortality Rate in Chengdu ZENGW ei, WEI ong-lan, LIAO Jiang C en terf or D isease C ontrol and Prev ention of Chengdu C ity, Chengdu 61004 1, Ch ina. AbstractObjectiveTo analyze and forecast the dynam ic trend of mortality rate in Chengdu by tmi e series. And to build a tmi e series model and evaluate its predictive effect. M ethodsACF and PACF analysis w ere used to identify themode.l Themodel parameterw as estmi ated by least squaremethod. Statistics ofBox-Ljung w as used to evaluate the degree of fitness ofAR IMA mode1, and the average relative errors of predictionw ere used as indexes to evaluate the predictive effect. ResultsThemultip le seasonalARIMA ( 0, 1, 1) ( 0, 1, 1) 12 model show ed the average relative errors reached 8150% . The morality rate descended gradually after 2000, and the forecast show ed that themortality ratew ou ld decrease continually in the next 2 years. Conclu- sion The fitness of tmi e series analysis and forecastmodel is good, so it can be used to analyze the regularity of disease or death variation and forecast its future tendency. Key wordsMortality rate; T mi e series; AR IMA model , ARIMA (p, , d, q ) (P, D, Q ) ,