New consistent characterization based goodness-of-fit tests

Abstract

We present new classes of consistent goodness-of-fit tests based on V-empirical Laplace transforms. The test statistics represent the quadratic and supremum-norm distances between appropriate functions. We inspect their properties for the particular case of testing null exponential distribution. In an extensive comparison study, we show that the tests have high Bahadur efficiencies. In addition, we compare them in the cases of small and moderate sample sizes via empirical powers.