目的:探讨时间序列模型中的自回归求和移动平均(ARIMA)乘积季节模型，预测成都市辐照悬浮红细胞临床供应量的应用。方法:选择2010年1月至2019年12月，成都市血液中心的辐照悬浮红细胞临床月供应量数据(n n＝108)为研究对象。其中，2010年1月至2019年6月的辐照悬浮红细胞临床月供应量数据(n n＝102)，用于ARIMA模型拟合；2019年7月至12月的临床月供应量数据(n n＝6)，用于验证所建立ARIMA模型的预测效果。采用Eviews 9.0统计学分析软件，对ARIMA模型进行拟合、参数估计及诊断。采用ARIMA模型预测2019年7月至12月成都市血液中心辐照悬浮红细胞的临床月供应量，并将预测值与实际值进行比较，从而对模型进行验证。n 结果:①根据2010年1月至2019年6月成都市血液中心的辐照悬浮红细胞临床月供应量数据，构建的时间序列模型为ARIMA(3，1，1)(1，1，1)n 12。该模型对成都市血液中心辐照悬浮红细胞临床月供应量的拟合良好[AR(1)＝-0.848、n t＝-4.340、n P<0.001，MR(1)＝-0.474、n t＝2.006、n P＝0.048，MR(2)＝-0.369、n t＝-2.723、n P＝0.008]，残差序列为白噪声(n Q＝22.24，n P＝0.163)。②采用ARIMA(3，1，1)(1，1，1)n 12模型，预测2019年7至12月成都市血液中心辐照悬浮红细胞临床月供应量的预测值，与其实际值的相对误差分别为5.02%、0.34%、4.98%、0.48%、1.26%和2.07%，平均相对误差为2.36%。n 结论:本研究建立的ARIMA(3，1，1)(1，1，1)n 12模型，能较好地拟合成都市血液中心辐照悬浮红细胞临床供应量变化趋势。该模型适用于辐照悬浮红细胞临床供应量的短期预测，并且为采供血机构辐照悬浮红细胞库存量管理提供依据。n “,”Objective:To explore the application of seasonal auto-regressive integrated moving average (ARIMA) model in forecasting the clinical supply of irradiated suspended red blood cells in Chengdu.Methods:The monthly clinical supply data (n n=108) of irradiated suspended red blood cells in Chengdu Blood Center from January 2010 to December 2019 were selected as subjects. Among them, the monthly clinical supply data (n n=102) of irradiated suspended red blood cells from January 2010 to June 2019 were used for ARIMA model fitting; and the data (n n=6) from July to December 2019 were used to verify the predictive effect of ARIMA model established in this study. The predictive amount predicted through ARIMA model and actual amount of monthly clinical supply of irradiated suspended red blood cells in Chengdu Blood Center from July to December 2019, were compared for further verifying the ARIMA model. Eviews9.0 statistical analysis software was used to fit, estimate and diagnose the ARIMA model.n Results:① Base on the monthly clinical supply data of irradiated suspended red blood cells in Chengdu Blood Center from January 2010 to June 2019, the time series model was ARIMA(3, 1, 1)(1, 1, 1)n 12. The ARIMA model fitted well to the monthly clinical supply of irradiated suspended red blood cells in Chengdu Blood Center [AR(1)=-0.848, n t=-4.340, n P<0.001; MR(1)=-0.474,n t=2.006, n P=0.048; MR(2)=-0.369, n t=-2.723, n P=0.008], and the residual sequence was white noise (n Q=22.24, n P=0.163). ② The relative error between the predictive amount which predicted by the ARIMA(3, 1, 1)(1, 1, 1)n 12 model, and the actual amount of monthly clinical supply of irradiated suspended red blood cells in the Chengdu Blood Center from July to December 2019, were 5.02%, 0.34%, 4.98%, 0.48%, 1.26% and 2.07%, respectively. And the average relative error was 2.36%.n Conclusions:The ARIMA (3, 1, 1)(1, 1, 1)n 12 model was well fit for the changing trend of clinical supply of irradiated suspended red blood cell in Chengdu Blood Center.This model is suitable for the forecasting of short-term clinical supply of irradiated suspended red blood cells, which could provide a basis for inventory management of irradiated suspended red blood cells in blood collection and supply institutions.n