Testing Gaussianity of Multivariate Data Using Entropy

Abstract

To date, many methods have been proposed for testing Gaussianity of a random process in both, the time and frequency domains. In this paper we analyse Gaussianity tests based on entropy for univariate and propose a new entropy based test for multivariate data. Although entropy-based tests for Gaussianity have not received much attention in the past, simulation results indicate that the tests are very competitive compared to other methods. Additionally, the entropy based test for Gaussianity of multivariate data has low computational cost which makes it attractive for practical applications.