Invariant tests for symmetry about an unspecified point based on the empirical characteristic function

This paper considers a flexible class of omnibus affine invariant tests for the hypothesis that a multivariate distribution is symmetric about an unspecified point. The test statistics are weighted integrals involving the imaginary part of the empirical characteristic function of suitably standardized given data, and they have an alternative representation in terms of an L²-distance of nonparametric kernel density estimators. Moreover, there is a connection with two measures of multivariate skewness. The tests are performed via a permutational procedure that conditions on the data.     

A characterization and a class of omnibus tests for the exponential distribution based on the empirical characteristic function

The characteristic function φ(t) of an exponentially distributed random variable is characterized by having its squared modulus identically equal to the real part of φ(t). We study the behavior of a class of consistent tests for exponentiality based on a weighted integral involving the empirical counterparts of these quantities, corresponding to suitably rescaled data. Bibliography: 25 titles.     

Two tests for multivariate normality based on the characteristic function

We present two tests for multivariate normality. The presented tests are based on the Lévy characterization of the normal distribution and on the BHEP tests. The tests are affine invariant and consistent. We obtain the asymptotic null distribution of the test statistics using some results about generalized one-sample U-statistics, which are of independent interest.      

Goodness-of-fit tests based on a robust measure of skewness

In this paper we propose several goodness-of-fit tests based on robust measures of skewness and tail weight. They can be seen as generalisations of the Jarque–Bera test (Bera and Jarque in Econ Lett 7:313–318, 1981) based on the classical skewness and kurtosis, and as an alternative to the approach of Moors et al. (Stat Neerl 50:417–430, 1996) using quantiles. The power values and the robustness properties of the different tests are investigated by means of simulations and applications on real data. We conclude that MC-LR, one of our proposed tests, shows the best overall power and that it is moderately influenced by outlying values.     

Goodness-of-fit tests based on Rao''s divergence under sparseness assumptions

In many practical situations the classical (fixed-cells) assumptions to test goodness-of-fit are inappropriate, and we consider an alternative set of assumptions, which we call sparseness assumptions. It is proved that, under general conditions, the proposed family of statistics based on Rao's divergence is asymptotically normal when the sample size n and the number of cells M_n tend to infinity so that n/M_n→ν>0. This result is extended to contiguous alternatives, and subsequently it is possible to find the asymptotically most efficient member of the family.     

Goodness-of-fit tests for copulas

This paper defines two distribution free goodness-of-fit test statistics for copulas. It states their asymptotic distributions under some composite parametric assumptions in an independent identically distributed framework. A short simulation study is provided to assess their power performances.     

Implementing empirical characteristic function procedures

The problem of estimating the first positive zero of the real part of a characteristic function is discussed. Knowledge of the location of this zero is essential for the application of inferential procedures based on the empirical characteristic function. A simple explicit nonparametric estimator is proposed and shown to solve simultaneously both the finite sample and asymptotic problems.     

A consistent modification of a test for independence based on the empirical characteristic function

A modification of a test for independence based on the empirical characteristic function is investigated. The initial test is not consistent in the general case. The modification makes the test always consistent and asymptotically distribution free. It is based on a special transformation of the data. Proceedings of the Seminar on Stability Problems for Stochastic Models, Moscow, Russia, 1996, Part I.     

Testing for independence by the empirical characteristic function

Tests of total independence of d (≥2) random variables are proposed using the empirical characteristic function. The approach is parallel to that of Hoeffding, Blum, Kiefer, and Rosenblatt.     

Goodness-of-fit tests for multivariate Laplace distributions

Consistent goodness-of-fit tests are proposed for symmetric and asymmetric multivariate Laplace distributions of arbitrary dimension. The test statistics are formulated following the Fourier-type approach of measuring the weighted discrepancy between the empirical and the theoretical characteristic function, and result in computationally convenient representations. For testing the symmetric Laplace distribution, and in the particular case of a Gaussian weight function, a limit value of these test statistics is obtained when this weight function approaches a Dirac delta function. Interestingly, this limit value is related to a couple of well-known measures of multivariate skewness. A Monte Carlo study is conducted in order to compare the new procedures with standard tests based on the empirical distribution function. A real data application is also included.     

A note on scale estimates based on the empirical characteristic function and their application to test for normality

We present a test for the composite hypothesis of normality against general alternatives. The test statistic is essentially the difference of two estimates of scale based on the empirical characteristic function. The test is simple to apply and its performance is comparable with that of several other well-known tests of normality.     

Statistical tests based on empirical processes and related questions

We consider goodness-of-fit tests for hypotheses about the forms of distributions and their membership in prescribed families of distributions. We first describe the classical tests based on empirical processes such as the omega-square tests of Cramér-von Mises-Smirnov and the Kolmogorov-Smirnov tests. We also consider Shapiro-Wilk tests. We devote a considerable amount of attention to testing the hypothesis that a random variable or vector is normal. We describe tests based on transformations of the empirical process, minimal distance tests and estimates, tests for symmetry, uniformity, and independence, and tests based on spacings. At the end we study methods of computing and the distribution functions of quadratic forms of normal random variables connected with tests of omega-square type. Bibliography: 372 titles.Translated fromItogi Nauki i Tekhniki, Seriya Teoriya Veroyatnostei, Matematicheskaya Statistika, Teoreticheskaya Kibernetika, Vol. 30, pp. 3–112, 1992.     

On nonparametric change point estimator based on empirical characteristic functions

We propose a nonparametric change point estimator in the distributions of a sequence of independent observations in terms of the test statistics given by Huˇskov′a and Meintanis(2006) that are based on weighted empirical characteristic functions. The weight function ω(t; a) under consideration includes the two weight functions from Huˇskov′a and Meintanis(2006) plus the weight function used by Matteson and James(2014),where a is a tuning parameter. Under the local alternative hypothesis, we establish the consistency, convergence rate, and asymptotic distribution of this change point estimator which is the maxima of a two-side Brownian motion with a drift. Since the performance of the change point estimator depends on a in use, we thus propose an algorithm for choosing an appropriate value of a, denoted by a_s which is also justified. Our simulation study shows that the change point estimate obtained by using a_s has a satisfactory performance. We also apply our method to a real dataset.     

The empirical characteristic function and large sample hypothesis testing

The empirical characteristic function is considered as a tool for large sample testing of a hypothesis that can be characterized in terms of the characteristic function. Two test statistics based upon the empirical characteristic function are proposed. The limiting distributions of these test statistics are obtained and methods are suggested for using these limiting distributions to calculate critical regions.     

A class of multiple sample tests based on empirical coverages

A class of multiple sample tests based on empirical coverages is proposed which is a generalization of Greenwood's and Sherman's one-sample goodness-of-fit test statistics. The asymptotic normality of the tests is established by embedding the empirical coverages into a stationary process satisfying the strong mixing condition.     

Goodness-of-fit tests for vector autoregressive models in time series

The paper proposes and studies some diagnostic tools for checking the goodness-of-fit of general parametric vector autoregressive models in time series. The resulted tests are asymptotically chi-squared under the null hypothesis and can detect the alternatives converging to the null at a parametric rate. The tests involve weight functions,which provides us with the flexibility to choose scores for enhancing power performance,especially under directional alternatives. When the alternatives are not directiona...