PDF (1939K)
摘要
目的 通过以经典时间序列模型 ARIMA 模型为基线,研究 1 套适用于弱季节性场景下的方法框架,将血液需求预测中的单纯模型精度比较转向具解释力的归因预测,为采供血垂类模型提供参考方案。方法 收集南宁地区 2018 年 1 月-2025 年 12 月的月度临床红细胞需求记录作为模型训练数据。 通过 STL 分解计算季节强度指数,采用 ADF 与 KPSS 检验联合判定序列的平稳性,然后分别构建 ARIMA、SARIMA、Holt-Winters、ARIMAX 及 ARIMA-XGBoost 混合模型,从模型结构、预测误差与业务解释力 3 个维度进行比较分析;采用动态时间规整(DTW)距离度量结合 K-means 对各年度月度序列进行形态聚类,分析弱季节性的年度差异,提出血液需求预测的方法论原则。结果 SARIMA 自动定阶为 ARIMA(4,1,0),Holt-Winters 季节平滑系数 γ= 1. 00,季节强度指数 FS = 0. 18,3 项检验均认为该序列为弱季节性时间序列。 DTW-KMeans 聚类(k = 3) 显示 2020-2022 年归为低值模式簇(年均约 27 002. 3 U),2018 年与 2023 年归为中值模式簇(年均约 38 755 U),2019 年与 2024-2025 年归为高值模式簇(年均约 44 284 U),说明弱季节性源于不同阶段的临床用血需求发生了变化。 样本外预测验证中,ARIMAX 模型表现最优,显著优于 ARIMA((DM 检验 P= 0. 011);时间序列交叉验证证实 ARIMAX 在最近两折中 RMSE、MAE、MAPE 3 项指标均最优,平均 RMSE(8 954. 46±3 530. 61) 和 MAPE(18. 75%±6. 83%) 标准差最小。 ARIMA-XGBoost 混合模型未表现出实质性精度提升(RMSE= 10 802. 20,MAPE= 21. 99%),劣于纯 ARIMA 模型。结论 数据序列呈现弱季节性波动,同时存在阶段性结构变化的场景下,融合业务解释变量的回归时序模型(ARIMAX)更具鲁棒性与管理价值的预测范式,提示血液需求预测应先验验证季节性假设、优先编码业务知识为外生变量、谨慎评估机器学习混合模型的适用边界。
Abstract
Objective To develop a forecasting framework suitable for weak-seasonality time series using the classical autoregressive integrated moving average (ARIMA) model as the baseline. Rather than focusing exclusively on predictive accuracy, we propose an interpretive, attribution-oriented approach to blood demand forecasting that emphasizes explanatory power over pure precision. The aim is to provide a domain-specific modeling reference for the blood collection and supply sector. Methods Monthly clinical red blood cell (RBC) demand data from Nanning, China (January 2018-December 2025) were collected. Seasonal strength was assessed by STL decomposition; stationarity was tested using ADF and KPSS procedures. ARIMA, seasonal ARIMA (SARIMA), Holt-Winters, ARIMAX, and a hybrid ARIMA-XGBoost model were built and compared in terms of model structure, forecast error, and operational interpretability. Annual monthly series were further subjected to DTW-based K-means morphological clustering to examine inter-annual variation under weak seasonality. Results SARIMA auto-selection resulted in an ARIMA(4,1,0) specification. The Holt-Winters seasonal smoothing parameter γ was 1. 00, and the seasonal strength index (FS) was 0. 18, confirming weak seasonality. DTW-KMeans clustering (k= 3) identified three phases:a low-demand cluster (2020-2022, annual mean≈27 002. 3 U), a medium-demand cluster (2018 and 2023, annual mean≈38 755 U), and a high-demand cluster (2019 and 2024-2025, annual mean≈44 284 U), suggesting that weak seasonality mainly reflects shifting clinical demand patterns across different stages. In out-of-sample validation, ARIMAX achieved the best performance (RMSE= 8 374. 88; MAE= 6 669. 88; MAPE= 14. 31%), significantly outperforming ARIMA (Diebold-Mariano test P = 0. 011). Time series cross-validation further confirmed that ARIMAX performed best in the last two folds, with the smallest standard deviations in RMSE (8,954. 46±3,530. 61) and MAPE (18. 75%±6. 83%), indicating superior predictive stability. The ARIMA-XGBoost hybrid showed no meaningful improvement (RMSE= 10,802. 20; MAPE = 21. 99%) and performed worse than standalone ARIMA. Conclusion For weakly seasonal blood demand series with structural change, ARIMAX-which explicitly incorporates explanatory variables-offers better stability and practical relevance than purely data-driven or hybrid machine learning alternatives. We recommend that blood demand forecasts test seasonal assumptions before model selection, encode domain knowledge as exogenous inputs where possible, and critically assess whether machine learning hybrids are warranted in low-seasonality settings.
关键词
Key words
[Author(id=1329502403779957549, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=35893139@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329502403842872111, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, authorId=1329502403779957549, language=EN, stringName=Wujin SU, firstName=Wujin, middleName=null, lastName=SU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Nanning Blood Center, Nanning 530007, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1329502403884815154, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, authorId=1329502403779957549, language=CN, stringName=苏武锦, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=南宁中心血站,广西 南宁 530007, bio={"content":"苏武锦(1980-),男,副主任技师,Email:35893139@qq.com
"}, bioImg=null, bioContent=苏武锦(1980-),男,副主任技师,Email:35893139@qq.com
, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329502403696071464, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, xref=null, ext=[AuthorCompanyExt(id=1329502403712848681, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, companyId=1329502403696071464, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Nanning Blood Center, Nanning 530007, China), AuthorCompanyExt(id=1329502403729625898, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, companyId=1329502403696071464, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=南宁中心血站,广西 南宁 530007)])]), Author(id=1329502403930952500, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329502403989672758, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, authorId=1329502403930952500, language=EN, stringName=Nan CHEN, firstName=Nan, middleName=null, lastName=CHEN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Nanning Blood Center, Nanning 530007, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1329502404031615800, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, authorId=1329502403930952500, language=CN, stringName=陈楠, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=南宁中心血站,广西 南宁 530007, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329502403696071464, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, xref=null, ext=[AuthorCompanyExt(id=1329502403712848681, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, companyId=1329502403696071464, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Nanning Blood Center, Nanning 530007, China), AuthorCompanyExt(id=1329502403729625898, tenantId=1045748351789510663, journalId=1155139928303341785, articleId=1329449373088756280, companyId=1329502403696071464, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=南宁中心血站,广西 南宁 530007)])])]
苏武锦,陈楠.
弱季节性视角下的采供血需求预测方法论-从竞争模型到归因范式的探讨[J].
中国输血杂志, 2026, 39(8): 1067-1073 DOI:10.13303/j.cjbt.issn.1004-549x.2026.08.011
| [1] |
杨旭,刘舒琦,黄慧,等. ARIMA 模型的临床红细胞需求预测研究[J].医学新知,2025,35(11):1272-1278.
|
| [2] |
YANG X, LIU S Q, HUANG H, et al. Clinical red blood cell demand prediction based on ARIMA model[J].New Medicine, 2025, 35(11):1272-1278.
|
| [3] |
谢淑红,张思静,严伟斌,等. 基于 ARIMA 模型的临床红细胞类血液需求预测研究[J].蚌埠医学院学报, 2023,48(5):633-636.
|
| [4] |
XIE S H, ZHANG S J, YAN W B, et al. Research on clinical demand prediction of red blood cell blood components based on ARIMA model[J].J Bengbu Med Coll, 2023, 48(5):633-636.
|
| [5] |
刘奇,陈红梅,罗川. 基于改进的蝗虫优化算法的红细胞供应预测方法[J].计算机科学,2021,48(2):224-230.
|
| [6] |
LIU Q, CHEN H M, LUO C. Red blood cell supply forecasting method based on improved grasshopper optimization algorithm[J].Comput Sci, 2021, 48(2):224-230.
|
| [7] |
倪茹玉,胡婉,张恒川,等. ARIMA 乘积季节模型与 LSTM 神经网络模型对我国麻疹发病数预测效果的比较[J].现代预防医学,2023,50(1):177-182.
|
| [8] |
NI R Y, HU W, ZHANG H C, et al. Comparison of predictive effect between ARIMA multiplicative seasonal model and LSTM neural network model on the number of measles cases in China[J].Modern Preventive Medicine, 2023, 50(1):177-182.
|
| [9] |
邱亚娟,张建平,罗佳,等. 三种时间序列模型预测红细胞类血液需求趋势效果对比[J].中国输血杂志, 2025,38(2):257-262.
|
| [10] |
QIU Y J, ZHANG J P, LUO J, et al. Comparison of three time series models in forecasting the demand trend of red blood cells[J].Chin J Blood Transfus, 2025, 38 (2):257-262.
|
| [11] |
DIEBOLD F X, MARIANO R S. Comparing predictive accuracy[J].Journal of Business & Economic Statistics, 1995, 13(3):253-263.
|
| [12] |
ZEBARI R R, ZEBARI G M, AL-ZEBARI A, et al. LSTM-XGBoost: an ensemble model for blood demand distribution forecasting-a case study in Zakho City, Kurdistan Region, Iraq[J].Oper Res Forum, 2025, 6 (1):1-22.
|
| [13] |
MIRI-MOGHADDAM E, KHOSRAVI BIZHAM S, MOEZZI F Z, et al. Long-term prediction of Iranian blood product supply using LSTM:a 5-year forecast[J].BMC Med Inform Decis Mak, 2024, 24(1):1-8.
|
| [14] |
MOTAMEDI M, DAWSON J, LI N, et al. Demand forecasting for platelet usage:from univariate time series to multivariable models[J].PLos One, 2024, 19(4):1-30.
|
| [15] |
冯晨,陈志德. 基于 XGBoost 和 LSTM 加权组合模型在销售预测的应用[J].计算机系统应用,2019,28(10): 226-232.
|
| [16] |
FENG C, CHEN Z D. Application of weighted combination model based on XGBoost and LSTM in sales forecasting[J].Computer Systems & Applications, 2019, 28 (10):226-232.
|
| [17] |
SCHAFFER A L, DOBBINS T A, PEARSON S A. Interrupted time series analysis using autoregressive integrated moving average (ARIMA) models:a guide for evaluating large-scale health interventions[J].BMC Med Res Methodol, 2021, 21(1):1-12.
|
基金资助
广西壮族自治区卫生健康委员会自筹经费科研课题(Z-A20251073)