IJSRP, Volume 3, Issue 7, July 2013 Edition [ISSN 2250-3153]
Rashpal Singh, Mohinder Pal
In this paper we formulate Stein–minimax type estimators and compare the performance properties with other estimators in case of mean value predictions. When the model is estimated by ordinary least squares it has been observed that least squares predicted is unbiased while minimax and Stein-minimax predictors are biased. The superiority conditions of the estimators have been derived by assuming error distribution to be non-normal.