基于时间序列模型的邯郸地区临床用红细胞分析及预测的初步研究  

Mathematical model based on time sequence of the handan region,the preliminary research on the clinical use red blood cell analysis and prediction

作　　者：孙国栋[1] 李俊霞[1] 陈慧[1] 李源[1] 王洪[1] 

机构地区：[1]邯郸市中心血站,河北邯郸056001

出　　处：《中国输血杂志》2015年第11期1361-1363,共3页Chinese Journal of Blood Transfusion

基　　金：河北省卫生和计划生育委员会医学科学研究重点课题计划;课题编号:ZD20140390

摘　　要：目的分析邯郸地区ABO血型系统各血型临床红细胞类用血分布规律,依据时间序列分析方法建立预测数学模型,并进行预测,指导血液机构的相关业务工作。方法对邯郸市2002-2013年每月各血型向临床供应红细胞类制品的量及合计血量进行建模。行Epidata3.0双录入数据,导入IBM SPSS Statistics 21,利用时间序列模型中专家建模器对各血型临床用血量及总血量建立数学模型,并预测2014年1-6月份的临床血液需求量,验证模型误差。结果专家建模器对红细胞类供血量给出的模型除了AB型为ARIMA(2,1,0)(0,1,1)以外,其余血型及总血量模型均为ARIMA(0,1,1)(0,1,1)模型。对5个模型残差的白噪声检验结果均显示P>0.05,说明残差均为白噪声序列,模型均提取了原序列中所有数据信息,模型诊断均得以通过。将预测结果与实际值进行比较,实际值均落入预测值的95%可信区间内,且平均相对误差较小,所得模型均为最优模型。结论通过建立数学模型的方式,补充后续数据,血液机构能够合理指导相应的业务工作,科学合理满足临床用血,规划用血趋势。Objective Analysing each ABO blood group in Handan area clinical blood transfusion distribution,red blood cells class based on time series analysis method to establish mathematical model for prediction and forecasting,guide blood agencies work related business. Methods Set up mathematical model for Handan city supply to clinical each month blood of red blood cells class each blood amount and total blood volume from January 2002 to December 2002. Line Epidata3. 0 double input data. Importing the IBM SPSS Statistics 21. Using time series expert modelers established clinical blood transfusion amount mathematical model,and predict 2014 1- 6 month demand for clinical blood,validation model error.Results Expert modelers to red blood cells in addition to the AB blood type for the given model of ARIMA（ 2,0）（ 0,1,1）,the rest of the blood and blood volume model of ARIMA（ 0,1,1）（ 0,1,1） model. Expert modelers in addition to the AB blood type for the given model of ARIMA（ 2,1,0）（ 0,1,1）,the rest of the blood type and the total blood volume model for ARIMA（ 0,1,1）（ 0,1,1）. The five models of residual white noise test results have shown the P〉 0. 05,instructions residuals all was white noise sequence,all models were extracted from the original sequence data information,model diagnosis were throughed. Comparison of the predicted results and actual values,the actual values are falling into a predictive value of 95% confidence interval,and the average relative error is small,the models all were optimal. Conclusion Through the establishment of mathematical model,add the follow-up data,blood agencies can reasonably guide the corresponding business work,scientific and reasonable to satisfy clinical use,planning trends in blood.

关 键 词：临床用血 ABO红细胞类制品 时间序列模型 

分 类 号：R457[医药卫生—治疗学] TP311.51[医药卫生—临床医学]