Additive and block decompositions of estimations under a general partitioned linear model

Professori  Tian vierailee matematiikan ja tilastotieteen laitoksella 11. - 15. helmikuuta.

Abstract

In this talk, I present a variety of results on additive and block decompositions of estimations under the following multiple partitioned linear model

M = {y, X1β1 + X2β2 + ⋯ + Xkβk, σ2Σ}

and its k small models

Mi = {y, Xiβi, σ2Σ},    i = 1,…,k.

In particular, I shall show how to derive identifying conditions for best linear unbiased estimators (BLUEs) under M to be the sum of the BLUEs under its k small models Mi.

Lisätietoja: lehtori Simo Puntanen, etunimi.sukunimi@uta.fi