Advances on Permutation Multivariate Analysis of Variance for big data

Abstract In many applications of the multivariate analyses variance, classic parametric solutions for testing hypotheses equality in population means or multisample and location problems might not be suitable various reasons. Multivariate lack a comparative study power behaviour most important combined permutation tests as number variables diverges. particular, it is useful to know under which conditions each different preferable terms power, how test increases when alternative hypothesis diverges, function proportion true hypotheses. The purpose this paper fill gap literature about tests, particular big data with large variables. A Monte Carlo simulation was carried out investigate application real case performed show utility method.