论文部分内容阅读
本文选用几项反映医院门诊和住院工作量指标,根据其逐年变化情况,用三次抛物线模型进行预测。结果表明,回代拟合时,模型对几组资料的拟合效果较好,R~2均大于0.8503,相对误差小于11.29％。外推预测时,误差明显增大,相对误差在2.05％～42.36％之间。当依事物的平均发展速度对未来某时刻进行估计,并将估计值与相应时间值作为一组对未来发展方向起控制作用的观察值参加模型建立时,则模型回代拟合效果仍较好,模型最小R~2为0.8554,而同时外推预测的精度也大为改善,相对误差在1.57％～16.55％之间。
This article selects several indicators that reflect the outpatient and inpatient workload of hospitals. Based on their year-to-year changes, three parabolic models are used for prediction. The results show that the model fitted well to several sets of data when fitting back, R~2 was greater than 0.8503, and the relative error was less than 11.29%. When extrapolating forecasts, the error increases significantly and the relative error is between 2.05% and 42.36%. When estimating some time in the future according to the average development speed of things, and taking the estimated value and the corresponding time value as a set of observations that control the future development direction to participate in the model establishment, the model backtracking fitting effect is still good. The minimum R~2 of the model is 0.8554, while the accuracy of the extrapolated prediction is also greatly improved. The relative error is between 1.57% and 16.55%.