基于SARIMA组合预测模型的血液供需预测研究

SARIMA-based Combined Forecasting Methods for Blood Supply and Demand

血液供应和需求在诸多因素的共同影响下,具有显著的不确定性,是实施有效的血液供应链管理所面临的巨大挑战。建立科学有效的血液供需预测模型,是努力实现血液供需匹配的前提和基础,以便在满足临床需求的同时尽可能减少血液资源浪费。为此,基于SARIMA模型分别建立了结合卡尔曼滤波的组合预测模型(K-SARIMA)和结合BP神经网络的组合预测模型(BPSARIMA),并根据MAPE、RMSE、RRSE等预测模型性能评价指标对所建模型进行了对比分析。实证分析结果表明,相较于SARIMA模型以及BP-SARIMA模型,采用K-SARIMA组合预测模型进行血液供需预测能够取得更好的预测效果。

Under the influence of various factors,there is a significant uncertainty in blood supply and demand,which is a great challenge to implement effective blood supply chain management.Establishing a scientific and effective model for forecasting blood supply and demand is the prerequisite and foundation of blood supply and demand matching,so as to reduce the waste of blood resources as much as possible while meeting clinical needs.Therefore,based on SARIMA,the K-SARIMA model combined Kalman filtering and the BP-SARIMA model combined BP neural network were built.Then the proposed forecasting models were compared according to the performance evaluation indices such as MAPE,RMSE,and RRSE.The empirical analysis results show K-SARIMA model has better performance in forecasting blood supply and demand.