Nonparametric hypothesis testing for equality of means on the simplex |
| |
Authors: | Michail Tsagris Simon Preston Andrew T. A. Wood |
| |
Affiliation: | 1. Department of Computer Science, Voutes, University of Crete, Heraklion, Cretemtsagris@yahoo.gr;3. School of Mathematical Sciences, Division of Statistics, University of Nottingham, Nottingham, UK |
| |
Abstract: | We investigate use of empirical and exponential empirical likelihood, and Hotelling and James statistics, to test the null hypothesis of equal population means based on two independent samples of data on the simplex. We perform an extensive numerical study using data simulated from various distributions on the simplex. The results, taken together with practical considerations regarding implementation, support the use of bootstrap-calibrated James statistic. |
| |
Keywords: | Compositional data hypothesis testing Hotelling test James test nonparametric empirical likelihood bootstrap |
|
|