ON ROBUSTNESS OF OPTIMAL LINEAR PREDICTION IN FINITE POPULATIONS

Jian-Ying Rong (P. R. China) and Xu-Qing Liu (P. R. China)

Received October 11, 2007

Abstract

This note is to deal with the problem of robustness of optimal linear prediction, which is an interesting and classical topic in the literature. We offer an alternative to the results derived by Xu and Wang [7], handle a simple case which is fulfilled in many situations, and apply the conclusions to a linear model with the regressor matrix having one column as the unit vector and the covariance matrix being compound symmetric. In addition, we make a further study of robustness from another perspective.

Keywords and phrases: robustness, linear unbiased prediction, finite population, compound symmetric, true (misspecified) model.