约束Welsch-Kuh统计量与约束Cook距离  

RESTRICTED WELSCH-KUH STATISTIC AND RESTRICTED COOK DISTANCE

作　　者：于义良[1] 吴诗泳 

机构地区：[1]山西师范大学

出　　处：《应用概率统计》1991年第2期136-142,共7页Chinese Journal of Applied Probability and Statistics

基　　金：山西省自然科学基金

摘　　要：本文讨论线性回归模型中约束条件下最小二乘估计的影响问题,利用约束下的Welsch-Kuh统计量研讨了剔除第i组数据(x＇_4,y_4)后,对x_i和x_y处拟合值的影响,另又用Cook统计量讨论了第i组数据对约束最小二乘估计的影响。This paper discusses the influence problem of least squares estimator for the linear regression model under the restrioted condition Using the W-K statistic under the restricted condition, we study the inflaences on the fitted values at x_i and x_i respectively while the i-th case (x＇_i, y_i) is deleted. The Cook's statistic is also used to discuss the i-th case's influence on least squares estimator under the restricted condition.

关 键 词：线性回归模型 W-K统计量 LS估计 

分 类 号：O212.1[理学—概率论与数理统计]