Econometrica: May 1978, Volume 46, Issue 3

Double k-Class Estimators of Coefficients in Linear Regression<705:DKEOCI>2.0.CO;2-1
p. 705-722

Aman Ullah, Shobha Ullah

This paper develops a new family of biased estimators, namely the double k-class, for the parameters of the general linear regression model. We note that James and Stein[7] Stein-rule estimator in the regression context is a member of the family of double k-class. The conditions for the existence of the moments and expressions for the exact and approximate bias, moment matrix, and the risk function of the double k-class estimator are analyzed.

Log In To View Full Content