利用温特斯加法指数平滑法预测某院出院人次

Prediction of the Number of Discharged Patients in a Hospital by Using Winters Additive Exponential Smoothing Model

目的利用温特斯加法指数平滑法建立某院出院人次预测模型,探讨其应用价值,为医院管理提供科学依据。方法采用SPSS21.0统计软件中的时间序列模块,用专家建模器对某院2005年1月至12月出院人次建立预测模型并预测2016年各月份出院人次,最后用2016年实际数据评价该模型的预测精度。结果残差为白噪声序列(P>0.05),模型提取了原时间序列中的所有数据信息,模型诊断通过,具有实际意义。预测结果显示,预测值均落在95%可信区间内,且与实际值相比平均相对误差均在9%以内,预测模型精度较高,预测效果良好。结论现阶段该院可以利用温特斯加法指数平滑法预测出院人次,合理安排工作,提高医院管理水平。

Objective The Winters additive exponential smoothing model is used to make prediction of the number of discharged patients in a hospital, and its application value is discussed to provide scientific basis for hospital management. Methods The time series module was used in SPSS21.0 statistical software. A prediction model was set up for the discharged patients in a hospital from January to December 2015 by using expert modeler, and forecast discharged patients each month in 2016, prediction accuracy evaluation of the model with the actual data in 2016. Results All of the residuals were white noise sequence（P〉0.05）. The model extracted all the data in the original sequence, and passed the model diagnosis, which has practical significance. The results shows that all the predicted value were within the 95% confidence interval, and the average relative error was less than 9%, compared with the actual value., so the predicting precision of the model was fine, and the prediction results were good. Conclusions At the present stage this hospital can use Winters additive exponential smoothing model to predict discharged patients and arrange work reasonably and improve hospital management.