用两种时间序列分析方法预测门诊量及效果评价
吴学智;何为虎;王安政
【期刊名称】《中国病案》 【年(卷),期】2016(017)008
【摘要】目的：对两种常见预测医院门诊量的方法进行探讨，找出适合某院的方法，为医院管理提供科学依据。方法分别用时间序列建立季节性指数平滑法模型和SARIMA模型，2015年数据用于验证所建立的模型，使用统计软件SPSS20.0完成。结果季节性指数平滑法模型预测总体相对误差为1.72%，各月预测数据相对误差均在10%以内；SARIMA模型预测总体相对误差为7.27%，有3个月的预测相对误差超过10%；季节性指数平滑法模型预测效果优于 SARIMA 模型。结论应选择预测误差较小的模型，现阶段该院宜选择季节性指数平滑法模型对门诊量进行预测，合理安排工作，提高医院管理水平。%Objectives To investigate the two common methods to predict outpatients’ number, find out the suitable method, and provide scientific basis for the management of the hospital. Methods The seasonal exponential smoothing model and SARIMA model were established by using time series. The 2015 data was used to validate the established models. Use statistical software SPSS20.0 to complete data processing. Results The overall relative error of the seasonal exponential smoothing model was 1.72%, and relative error of each month forecast data was less than 10%;The overall relative error of SARIMA model prediction was 7.27%, there are 3 months of forecast relative error were
用两种时间序列分析方法预测门诊量及效果评价
