An angular approach for linear data

Abstract

A procedure for estimating the slope in a linear model which was introduced by Chambers & Heathcote (1981) is reexamined. A careful analysis of the procedure leads to a geometric interpretation in terms of angular random variables. This interpretation motivates a method of estimating the location or intercept, simplifies the evaluation of the robustness of the procedure and provides insight into the relationship between the method and other robust techniques. Show more