Linear sufficiency with respect to a given vector of parametric functions (Q1819867): Difference between revisions

Let \(\{\) Y,X\(\beta\),V\(\}\) be a general Gauss-Markov model. FY is said to be linearly sufficient for the estimable function \(K\beta\) if the minimum dispersion linear unbiased estimator of \(K\beta\) is a linear function of FY. Theorem 1 gives a necessary and sufficient condition for such a linear sufficiency. Corollaries deal with some special cases, while theorem 2 considers the problem of linear minimal sufficiency. Applications are given to models with additional information on nuisance parameters, stepwise estimation and seemingly unrelated regressions.