Testing exchangeability of multivariate distributions

Although there have been a number of available tests bivariate exchangeability, i.e. symmetry for distributions, the literature is void whether multivariate distribution with more than two dimensions exchangeable or not. In this paper, permutation exchangeability distributions are proposed, which based on non-parametric combination methodology, combining tests. Numerical experiments real as well simulated data presented here. The test turns out to be typically powerful performed only over single pair variables, and also suitable compared exploiting approaches Benjamini–Yekutieli Bonferroni.