ABSTRACT
In this Chapter in Section 3.1 definition of the general linear model is presented. In Section 3.2 estimates of regression coefficients are derived first using least squares and then the maximum likelihood method. Then in Section 3.3 the likelihood ratio test of linear hypotheses on regression coefficients is considered. In Section 3.4 various profile likelihood-based confidence procedures are discussed. These include joint confidence regions for finite sets of linear combinations of regression coefficients and confidence intervals for linear combinations of regression coefficients. Finally, in Section 3.4 model criticism in terms of residuals is discussed.
