
Title
A Simpler, Affine-Invariant, Multivariate, Distribution-Free Sign Test
Speaker
Ronald H. Randles, Department of Statistics, University of Florida
Abstract
A simpler multivariate sign test is proposed which uses the transformation retransformation approach of Chakraborty, Chaudhuri and Oja (1998) together with a directional transformation due to Tyler (1987). This produces a multivariate sign test that is practical to apply to data of any dimension, makes minimal assumptions about the underlying distribution and has a small sample distribution-free property over a broad class of population models. It is shown to perform very well in comparison to Hotelling's T^2 and other multivariate sign tests for heavy-tailed and skewed distributions.