唐美燕,张玉芳,吴小莉

本文基于最小二乘原理，结合舍入误差的均匀分布与独立假设，探究四舍五入舍入数据对一元线性回归系数的影响机制。研究发现，四舍五入舍入数据会导致回归斜率产生系统性向下偏倚，截距偏差由斜率偏差传导且受多因素调节；该偏倚为系统性固定偏差，无法通过增大样本量消除。据此，进一步给出舍入精度值设定、模型与预测结果校正的应用路径，为实际建模中舍入数据的回归分析提供理论参考与应用指导。

最小二乘原理；舍入误差；回归系数；模型校正

[1] 胡果荣.基于舍入数据的统计推断[D].吉林大学,2006. [2] 宋扬.四舍五入数据的线性回归问题[D].清华大学,2017. [3] Sheppard W F.On the calculation of the most probable values of frequency constants for data arranged according to equidistant divisions of a scale[M].Proc.london math.Soc,1898,29:353-380. [4] Schneeweiss H,Komlos J,Ahmad A S.Symmetric and asymmetric rounding:a review and some new results[J].AStA Adv StatAnal,2010,94:247-271. [5] Heitjan D F,Rubin D B.Ignorability and coarse data[J].Ann Statist,1991,19:2244-2253. [6] Hall P.The influence of rounding errors on some nonparametric estimators of a density and its derivatives[J].Sima APPl math,1982,42:390-399.

点击下载PDF