The American Statistician, November 1st, 1999
We present an approach to the problem of general least squares estimation of the general linear model in terms of constrained optimization, which is in turn solved via Lagrange multipliers. We demonstrate that one system of equations is sufficiently versatile to cover not only the estimation of new observations, of fixed parameters in regression and of fixed and random effects in mixed models, but also of the diagnostics associated with conditional and marginal residuals and of subset deletion.
KEY WORDS: Best linear unbiased estimation; Best linear unbiased prediction; Conditional residual...
HighBeam Research, Free Preview: 'Simplifying General Least Squares.'... Full Membership required for unlimited access. Free 7-day trial.
Subscribers: HighBeam content is only available to HighBeam subscribers. Click the link above for more information.
