Tests of linear hypotheses based on indicators

A method of linear hypothesis testing isdeveloped and investigated for a linear model with independent random errors that have quantiles of a given levels specified by unknownparameters. This kind of prior information allows describing an unknown scale of distribution through the interquantile range. A number ofmulti-sample problems are considered; indicator-based tests for homogeneity of samples by locations and/or scales are derived.

Keywords

Authors

References