Inference on Parameters in a Linear Model a Review of Recent Results.
Abstract
This paper, in three parts, is a review of recent results on inference on parameters in a linear model. In the first part, the Gauss-Markoff theory is extended to the case when the dispersion matrix of the observable random vector is singular. In the second, robustness of inference procedures for departures in the design matrix, the dispersion matrix and distributional assumptions about the error components is considered. Finally, the third part introduces concepts of linear sufficiency and completeness in linear models, without making any distributional assumptions. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1983
- Accession Number
- ADA127486
Entities
People
- Bimal Kumar Sinha
- Calyampudi Radhakrishna Rao
- Jochen Muller
Organizations
- University of Pittsburgh