Inferential procedures based on the integrated empirical characteristic function

Abstract

We introduce the novel notion of the integrated characteristic function and its empirical counterpart. Some basic properties of these new objects are mentioned and in turn utilized in order to construct new procedures for testing goodness of fit to parametric distributions, for testing symmetry and homogeneity, and for testing independence. Asymptotic results are obtained, while corresponding Monte Carlo results on the finite-sample behavior of the procedures are also included