时间序列ARIMA模型在医院剖宫产统计中的预测研究

Research on prediction of time series ARIMA modeling in caesarean section

目的建立医院剖宫产统计的预测模型,为医院决策提供依据。方法根据湖南省妇幼保健院2005~2012年产房剖宫产统计的数据,运用时间序列ARIMA模型预测分析医院未来两年剖宫产的发展趋势。结果 2009年之后医院的数据波动很频繁,上升趋势较强,最终利用2009年以后的原始数据序列进行分析建模。建立ARIMA(1,1,1)(0,1,1)模型,模型方程为▽▽12Xt=(1+0.705B)(1+0.705B12)/(1+0.874B)×at。结论 ARIMA(1,1,1)(0,1,1)模型预测未来两年医院剖宫产趋势呈明显的季节性、周期性,7~9月呈现出高峰点,10~12月呈现全年的最高峰值。结论本预测模型可为决策者对医院政策调整提供科学、合理的依据。

Objective To build the caesarean prediction modeling and provide the evidence for the hospital. Methods The devel-opment trend of caesarean in the next two years was analysed according to the caesarean data in MCH hospital of Hunan province from 2005 to 2012. Results Since 2009, volatile data showed raising trend. Therefore, the modeling was well done by original data sequence. ARIMA modeling (1,1,1) (0,1,1) was built, and model equation was▽▽12Xt =(1+0. 705B)(1+0. 705B12)/(1+0. 874B) × at. Conclusion ARIMA modeling (1, 1, 1) (0, 1, 1) predicts that the datum of caesarean shows obviously seasonal and periodic difference. The differ-ence presents the peak from July to September but reaches the top from October to December. The modeling provides the scientific and rea-sonable evidence for the hospital management.