Consistent Directions of the Least Squares Estimators in Linear Models.
Abstract
The consistent directions of the least squares estimators in a linear model are defined to be the linear combinations of parameter estimates that are asympotically consistent. When the design variable is univariate and the regression function is smooth, consistent directions are characterized in previous papers (Wu, 1980; Wu and Wang, 1982) in terms of the convergence rates of the design sequence to its limit points. Extension of these results to multivariate design variables are considered in the present paper.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1982
- Accession Number
- ADA120992
Entities
People
- C. F. Jeff Wu
- Song-gui Wang
Organizations
- University of Wisconsin–Madison