时间序列在辐照单采血小板供应管理上的应用分析  

Application of time series ARIMA model in the distribution management of irradiated apheresis platelets

作　　者：张永鹏[1] 席光湘[1] 洪缨[1] 万浬科[1] 陈迎春[1] 赵玉伟[1] ZHANG Yongpeng;XI Guangxiang;HONG Ying;WAN Like;CHEN Yingchun(Chengdu Blood Center,Chengdu 610041,China)

机构地区：[1]成都市血液中心,四川成都610041

出　　处：《中国输血杂志》2020年第9期975-977,共3页Chinese Journal of Blood Transfusion

基　　金：成都市医学科研课题(2017040)。

摘　　要：目的根据成都地区辐照单采血小板临床月使用量,应用时间序列ARIMA模型预测临床辐照单采血小板月使用量,为单采血小板采集和献血者管理工作提供科学的参考依据。方法收集2010年1月—2019年6月辐照单采血小板的月供应量,用eviews9.0统计分析软件对该时间段的临床用量数据建立ARIMA模型,预测2019年7—12月成都地区临床辐照单采血小板的月供应量,与实际值进行比较,验证模型。结果临床辐照单采血小板供应量模型构建为ARIMA(1,1,1)×(1,1,1)12,2019年7—12月相对误差分别为1.08%,4.49%,0.98%,2.80%,3.84%和5.80%;预测平均相对误差为3.17%,相对误差均基本上在6%以内,预测效果较好。结论 ARIMA模型能较好地拟合成都地区临床辐照血小板供应量变化趋势,适用于临床辐照血小板供应量的短期预测,为血小板采集和献血者管理提供依据。Objective To analyze the monthly irradiated apheresis platelets(IAPs) distribution currently in Chengdu to predict future monthly IAP demand using time series ARIMA model,so as to provide scientific basis for platelet collection and blood donor management.Methods Data relative to monthly distribution of IAPs from January 2010 to June 2019 was reviewed, and ARIMA model was established based on this period using Eviews 9.0 statistical analysis software then used to predict future monthly IAPs distribution(July to December 2019). The model was verified in comparison to the actual values. Results The IAPs distribution model was constructed as ARIMA(1,1,1)×(1,1,1)12.The relative errors from July to December 2019 were 1.08%, 4.49%, 0.98%, 2.80%, 3.84% and 5.80%, respectively. The relative errors were almost within 6%,with an average relative error of prediction at 3.17%, suggesting a good prediction effect. Conclusion The ARIMA model is very informative in exhibiting distribution trend of IAPs in Chengdu and suitable for short-term prediction, which helps in decision making for platelet collection and blood donor management.

关 键 词：辐照单采血小板 ARIMA模型 供应量预测 

分 类 号：R331.143[医药卫生—人体生理学] R457.12[医药卫生—基础医学]