A TEST FOR SYMMETRY BASED ON DENSITY ESTIMATES

Abstract

This paper describes a test for symmetry based on density estimates. We use a kernel typed estimator as a density estimator. We propose the statistic based on density estimates, and derive its asymptotic distributions under both the null hypothesis and the alternative. Furthermore, we give the asymptotic power and we compare it with other methods. We illustrate that our method is better in some cases both theoretically and practically.