Simultaneous confidence procedures in multivariate calibration problems

Abstract

In a multivariate linear regression model simultaneous confidence procedures for some of the independent variables are constructed. Two of these procedures are extensions of similar methods which were first presented by R. G. Miller Jr. They are based on the so-called Bonferroni inequality and the union intersection principle, respectively. Furthermore a third method (which is only applicable in an univariate model) is presented. This procedure is an improvement of one of the two methods given by Miller.