Non-Parametric Approach of Estimation in Linear Models

 

K.K. Saxena and O.P. Srivastava (1989). Non-Parametric Approach of Estimation in Linear Models. Journal of Statistical Research, Vol. 23, No. 1-2, pp.  ??-??.

 

Abstract

An estimator of $ \beta $ in the model $ y_i=\alpha+\beta x_i+e_i $, where $ e_i $ are not necessarily normally distributed, is proposed by using the means of all possible slopes. Its relation with U-statistics and jackknife has been shown. It has also been shown that by using jackknife technique, we get an estimate of variance of estimator directly from pseudo-values. This is one of the main advantages over other existing distribution free estimators. A special case, where $ x_i $'s are equidistant is also studied and the efficiency of this estimator is obtained in comparison with the ordinary least squares estimators.

 

Fulltext PDF not available for this article.