ARIMA模型应用于临床悬浮红细胞用量预测的探讨

Forecast of clinical demand of RBC in a hospital based on ARIMA

目的建立分析和预测临床用血量的回归滑动平均混合模型（ARIMA）,为科学预测临床血液用量及合理储血提供依据。方法使用SPSS16.0软件做模型的识别、模型的参数估计与检验,建立ARIMA模型,对本院2007~2012年临床悬浮红细胞用量进行统计分析,将2013年预测值与实际值比较,检验模型的预测能力。结果构建ARIMA（0,1,1）（0,1,0）12模型,残差序列是白噪声,拟合优度相对较好,该模型的平均百分误差为5.78%,预测效果良好。结论 ARIMA模型能够科学、有效、客观反映时间序列的变化规律,为以后医院临床用血及合理储血提供了依据。

Objective To develope a autoregressive integrated moving average model, for statistical analysis of the clini- cal demand of red blood cell（RBC） over the past five years ina hospital, so as to provide evidence forl scientific prediction of the amount of blood for clinical use and reasonable blood reserve. Methods SPSS16.0 software was used for model identifi- cation, parameter estimation and test. so as to establish a mixed regression moving average model （ARIMA）. ARIMA model was applied to analyze the data of clinical RBC use obtained from the years 2007 to 2012 and forecast the seasonal RBC de- mand of 2013 in our hospital. Then the predicted value was compared with the actual value of 2013 to test the forecastinz abil-ity of the model. Results The ARIMA（0,1,1） （0,1,0） 12 model was perfectly matched. The average relative error was 5. 78%. The model had a good predictive effect. Conclusion ARIMA model can reflect the changing rule of time sequence sci-entifically, effectively and objectively. It can provide strong basis for clinical blood use and reasonable blood reserve.