A1287
Title: Testing central symmetry: A data depth approach
Authors: Bojana Milosevic - University of Belgrade (Serbia) [presenting]
Jana Zivkovic - University of Belgrade (Serbia)
Abstract: A class of nonparametric tests for central symmetry in multivariate distributions based on statistical data depth is proposed. The approach relies on a depth-based transformation of the sample, enabling the use of classical one-dimensional test statistics within a unified framework. Several test statistics, including Kolmogorov-Smirnov, Cramer-von Mises, Anderson-Darling, and energy distance-based measures, are considered. A bootstrap calibration procedure is developed to ensure accurate finite-sample performance. Simulation studies demonstrate that the proposed methods maintain the nominal level and exhibit strong power across a variety of asymmetric distributions, often outperforming existing competitors.
