# Talk:Method of least squares

This article contains actually a short description of *linear regression*, a subject described more extensively in linear regression.

I would make this a more general introduction to general least squares adjustment.

The article titled "least squares" now has links to linear regression and Gauss-Markov theorem, both of which are fairly good. Note that the linear regression article states that the model

where the last term is the only one to the right of "=" that is random, and those are i.i.d. normally distributed with expectation 0, does not cease to be **linear** because of the presence of the quadratic term, precisely because the vector least-squares estimates of the three coefficients is linear in the vector of *Y*-values. I mention this because the language that was here before I redirected this article makes me wonder if that was appreciated. Michael Hardy 17:32 Mar 3, 2003 (UTC)