Title: On the BHEP test for functional data
Authors: Maria Dolores Jimenez-Gamero - Universidad de Sevilla (Spain) [presenting]
Norbert Henze - Karlsruhe institute of Technology (Germany)
Abstract: Many statistical procedures for finite dimensional data assume the data To be normally distributed. Thus, a number of normality tests have been proposed. Some of them are based on comparing a nonparametric estimator of a function characterizing a probability law with a parametric estimator of that function, obtained under the null hypothesis. The BHEP test belongs to this class of tests. It is based on comparing the empirical characteristic function with the characteristic function of the normal law. The normality assumption is important not only in the classical context, but also in other settings such as functional data analysis. Since the probability measure of a random element taking values in a separable Banach space is characterized by its characteristic function, the objective is to extend the BHEP to the functional data context.