广义相加模型在上海世博会园区医疗站就诊人数预测中的应用
Application of Generalized Additive Model in Forecasting Cases of Medical Sites in Expo 2010 Shanghai China
-
摘要:
目的 探讨应用广义相加模型进行中国2010年上海世博会园区医疗站就诊人数预测的可行性。 方法 采用时间序列的广义相加模型,在控制星期效应的基础上,对2010年5月1日至8月8日开园100 d间的每日世博园区就诊人数、入园人数和气象因素资料进行模型拟合。 结果 园区就诊人数存在周末效应,就诊人数随入园总人数的增加而增加;风速对就诊人数的影响趋势是随风速的增加,就诊人数先增加后减少;随着气温上升和日温差增加,就诊人数呈上升趋势(P < 0.01)。利用模型对每日就诊人数进行预测,预测的平均相对误差为10.44%。 结论 广义相加模型能较好拟合世博园区医疗站就诊人数的趋势,可用于大型活动中的预测研究。 Abstract:Objective To explore application of generalized additive model in forecasting cases of medical sites in Expo 2010 Shanghai China. Methods By using a generalized additive model with time series adjustment for weekend, the daily cases of medical sites, the total visitors to the Expo and the meteorological factors from May 1st to Aug 8th were analyzed and the forecast was compared with the actual cases. Results The number of medical site cases showed a weekend effect. With the total visitors increasing, the number of cases increased first and then decreased. With the daily temperature increased, the number of cases increases. The average relative error of prediction was 10.44%. Conclusion A generalized additive model fit the number of cases of the Expo medical stations and can be used for similar forecast in mass gathering in the future.
下载:
