数据的剔除对Minimax线性估计的影响与相关系数  

The Influence of Data-Deleting on Mininmax Linear Estimation and Correlation Coefficients

作　　者：田保光[1] 王建新[1] 翟发辉[1] 

机构地区：[1]青岛科技大学数理系,山东青岛266042

出　　处：《四川师范大学学报（自然科学版）》2007年第6期714-716,共3页Journal of Sichuan Normal University（Natural Science）

摘　　要：用方差比和广义方差比作为判定数据对Minimax线性估计影响的标准,建立了剔除一组数据时的方差比与相关系数的关系,找到了剔除多组数据时的广义方差比与Hotelling广义相关系数的联系,并讨论了剔除一组数据与剔除二组数据时的广义方差比之间的关系.The variance ratio and generalized variance ratio are used as criterion of influence of data on minimax linear estimation. The relationship between variance ratio and correlation coefficient is built when one data case is deleted. The relationship between generalized variance ratio and Hotelling generalized correlation coefficient is ＇also found when a set of data cases is deleted. Moreover, the relationship between generalized variance ratioes when one data case is deleted or two data cases are deleted is also discussed.

关 键 词：Minimax线性估计 方差比 广义方差比 相关系数 Hotelling广义相关系数 

分 类 号：O212.1[理学—概率论与数理统计]