Sankhya: The Indian Journal of Statistics

2001, Volume 63, Series B, Pt. 1, pp. 10--26

OPTIMAL DATA AUGMENTATION FOR THE ESTIMATION OF A LINEAR PARAMETRIC FUNCTION IN LINEAR MODELS

By

DULAL KUMAR BHAUMIK University of South Alabama, Mobile, U.S.A.

and

THOMAS MATHEW University of Maryland, Baltimore County, U.S.A.

SUMMARY. In the setup of a linear regression model, the problem of augmenting a given set of observations is investigated, when the inference problem is the estimation of a linear parametric function of the mean vector. For this problem, the optimal selection of additional observations is studied. The optimal design matrix is constructed following the A, D and E-optimality criteria. The results are illustrated with an example.

AMS (1991) subject classification. Primary 62K05; secondary 62J05.

Key words and phrases. A-optimality, D-optimality, E-optimality, selection of additional observations.

Full paper (PDF)