应用趋势季节模型预测医院门诊人次

Seasonal Trend Model in Forecasting Hospital Outpatient Visits

目的建立趋势季节模型,预测某医院2011年各季度门诊人次。方法采用移动平均法从原时间序列中分离季节变动、不规则变动、长期趋势和循环变动的序列,并使用合适计算方法得到各个趋势季节模型的不同指数因子,同时采用预测误差(SE)计算预测值范围。结果某医院2011年的各个季度的门诊人次预测值范围为144 049～164 7681、47 150～167 8681、78 660～199 3781、80 005～200 723。结论趋势季节模型的适用条件为在时间序列中既有长期趋势,又存在季节变动,且必须有连续不少于4个周期的资料,同时要选择客观现象比较稳定的周期。

Objective To establish a seasonal trend model to forecast the hospital outpatient visits in different seasons in 2011.Methods Seasonal changes,irregular changes,long-term trends and cyclic changes were separated from the original time series by the moving average method.Different factor indexes of seasonal trend model were obtained by using the appropriate model,and predictive ranges were calculated by predictive standard errors（SE）.Results The predictive ranges of hospital outpatient visits of each season in 2011 were 144 049-164 768,147 150-167 868,178 660-199 378,and 180 005-200 723.Conclusion The applicable conditions of the seasonal trend model are as follows： the time series includes both the long-term trends and seasonal variation;The data of time series of more than four continuous cycles must be included and the steady periods of objective phenomena must be selected.