论文部分内容阅读
目的:分析宜昌地区ABO血型系统各血型红细胞用量的分布规律,依据时间序列分析方法建立自回归积分滑动平均模型(ARIMA)进行预测,指导采供血机构相关业务工作。方法:在SPSS18.0中利用时间序列模型中专家建模器,对宜昌市2008-01-2015-12红细胞总的用量及各血型分别用量建立数学模型,并预测2016年1至6月用量,与实际用量对比,验证模型误差。结果:专家建模器对红细胞总量、A型及O型红细胞用量给出的模型是ARIMA(0,1,1)(0,1,1),B型和AB型红细胞用量给出的模型分别是ARIMA(1,1,1)(1,1,1)和ARIMA(2,1,1)(1,1,1)。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,模型提取了原序列中所有数据信息,模型诊断得以通过。将预测值与实际值进行比较,实际值均落入预测值95%的可信区间内,且平均误差相对较小,模型预测效果良好。结论:ARIMA模型能够科学、有效地反映时间序列的变化规律,可以有效预测短期红细胞用量,有针对性地指导血站的采供血业务工作。
OBJECTIVE: To analyze the distribution of erythrocytes in blood of ABO blood group system in Yichang, and to establish the ARIMA model based on time series analysis to predict the related business of blood collection and blood supply agencies. Methods: Based on SPSS18.0, an expert modeler was used to establish the mathematical model of the total amount of erythrocytes and the amount of each blood type in Yichang City from January 2008 to December 2015, Contrast with actual dosage, verify model error. Results: The model given by expert modeler for the total amount of erythrocytes, type A and type O erythrocytes was the model given by ARIMA (0,1,1) (0,1,1), type B and type AB erythrocytes ARIMA (1,1,1) (1,1,1) and ARIMA (2,1,1) (1,1,1), respectively. The results of white noise test on 5 models showed P> 0.05, indicating that the residuals were all white noise. All the data in the original sequence were extracted and the model diagnosis was passed. The predicted value is compared with the actual value, the actual value falls within the 95% confidence interval of the predicted value, and the average error is relatively small, and the model prediction effect is good. Conclusion: The ARIMA model can reflect the change rule of time series scientifically and effectively, and can effectively predict the amount of short-term erythrocytes and guide the blood collection and blood supply work in a specific way.