5 Matrix Algebra and Quadratic Forms

From here on, the model has \(k\) regressors and is written in matrix form. This section collects the matrix results and the distribution theory that the rest of the course uses. Most of it is the theory of quadratic forms \(X'AX\) in normal random vectors, which is what lies behind every \(\chi ^2\), \(t\) and \(F\) statistic in regression. At the end the theory is applied to the residuals of a regression, which delivers the two facts that Section 2.4 had to take on trust.

Questions on this section

Stuck on something here? Ask below and it stays attached to this topic.