时间序列分析法在献血招募工作中的应用

Application of time series analysis method in blood center

目的预测下一年度的献血人次数等指标,据此制定下一年度献血者招募计划。方法将2003年1月-2012年12月武汉血液中心献血人次数据作为样本数据,以月为维度,用时间序列法中的季节性相加模型作为预测方法,用R软件来实现预测,并用2013年的数据作为检验数据对预测的准确性做验证。结果计算得到2013全年的全血献血人次数预测值为136396,与实际值144884相比减少了5.86%。实际数据均在模型的95%置信区间内,预测合理性检验为合理。结论在没有大的政策变动等因素影响时,时间序列预测法-季节性相加模型可以快速简便地对下一年度的全血献血人数做出基本的预测。以此类推,该方法还可以应用在成分献血人次数的预测,供血量的预测等方面。

Objective To predict the number of donors next year and other indicators and to formulate a plan on donor recruitment next year. Methods Samples were obtained from the data from January 2003 to December 2012. Months were used as an independent dimension. The additive model with a seasonal component was used as a prediction method in the time series analysis. R software was used to make the prediction. Lastly,the accuracy of the data from 2013 were verified. Results The number of whole blood donors was predicted to be 136,396. This number was reduced by 5.86% ,when compared with the actual value of 144 884. The actual data were within the 95% confidence interval of the model. The prediction was tested to be reasonable. Conclusion In the absence of prominent changes in policies, the time series method could quickly and eas- ily predict the number of whole blood donors in the upcoming year. The time series analysis method can also be applied in the prediction on the number of components donated and the amount of blood supply.