论文部分内容阅读
目的:在移动平均趋势剔除法、最小二乘法两种趋势剔除法中找出一种能较好反映某医院门诊量季节变化规律的方法。方法:根据该医院2005~2008年各月门诊数据,运用移动平均趋势剔除法、最小二乘法分别计算该医院门诊患者的季节指数(seasonal index)和各月预测值,并对预测结果进行平均绝对偏差(MAD)、平均平方误差(MSE)、平均预测误差(AFE)、平均绝对百分误差(MAPE)分析。同时,判断实际值与预测值的容许区间的关系。结果:移动平均趋势剔除法和最小二乘法预测值的MAD、MSE、AFE、MAPE分别为766.94,888236.8542,-0.23,5.4782498%和739.0196,802281.2,-0.125,5.259 453%。移动平均趋势剔除法有4个实际值落在容许区间之外,最小二乘法有2个。结论:最小二乘法能够更好反映出该院门诊量季节变化的规律,是预测的最佳选择方案。
OBJECTIVE: To find out a method that can better reflect the seasonal variation of outpatient service volume in a hospital from the trend elimination method and the least square method. Methods: According to the outpatient data from 2005 to 2008 in each hospital of the hospital, the seasonal index and monthly forecast of outpatients in the hospital were calculated respectively by the moving average trend elimination method and the least squares method. The average absolute MAD, MSE, AFE, and MAPE analysis. At the same time, determine the relationship between the actual value and the allowable range of the predicted value. Results: The MAD, MSE, AFE and MAPE of the moving average trend subtraction method and the least square method were 766.94,888236.8542, -0.23,5.4782498% and 739.0196,802281.2, -0.125,5.259 453% respectively. Moving average trend culling method has four actual values fall outside the allowable range, the least squares method has two. Conclusion: The least squares method can better reflect the seasonal variation of outpatient service in this hospital and is the best choice for prediction.