ABSTRACT
The simplest version of the linear model arises with one of the basic problems in elementary statistics,where yi are randomly sampled fromapopulationwith unknown mean μ and variance σ 2. In this case, Xb takes a very simple form
Xb = 1(μ)
so that the unknown coefficient vector b is just the scalar μ. This very simple model will be extended to handle cases such as heteroskedasticity inChapter 4 ormultivariate responses in Chapter 9.
