Goodness-of-fit tests for a multivariate distribution by the empirical characteristic function (Q1365549)

For testing the goodness of fit of a distribution function to a given one, based on a random sample of size \(n\), the author considered a quadratic form of order \(2m\) composed from the deviations of the \(m\) real part and the \(m\) imaginary part of the empirical characteristic function from their expected values measured at \(m\) given points. Then the asymptotic null distribution of the test statistic is obtained. It is shown to be a normal distribution or weighted sum of independent chi-square distributions with one degree of freedom, according to the choice of the weight matrix of the quadratic form and the limiting condition on \(m\) as \(n\) tends to infinity. An extension to a composite null hypothesis is considered using a consistent estimator.