No label defined (Q4246928)

Let \(\beta_n^U\) and \(\gamma_n^U\) denote the uniform empirical process and the uniform quantile process of sample size \(n\), respectively. The uniform Bahadur-Kiefer process \(\beta_n^U + \gamma_n^U\) is a well-investigated tool in asymptotic theory which can be used to derive the asymptotic properties of functionals of the `complicated' quantile process \(\gamma_n^U\) from the asymptotic properties of functionals of the `easier' empirical process \(\beta_n^U\). \textit{W. Vervaat} [Z. Wahrscheinlichkeitstheor. Verw. Geb. 23, 245-253 (1972; Zbl 0238.60018] initiated the study of the integrated Bahadur-Kiefer-process \[ V_n^U(t) = \int\limits_0^t\{ \beta_n^U(s) + \gamma_n^U(s)\} ds, \qquad 0 \leq t \leq 1 , \] also called integrated empirical difference process. A generalization of \(V_n^U\) to arbitrary distribution functions \(F\) is \[ V_n(t) = \int\limits_{-\infty}^{F^{-1}(t)} \beta_n(x)dx + \int\limits_0^t \gamma_n(s)ds, \qquad 0 \leq t \leq 1 , \] where \(\beta_n = F_n - F\) and \(\gamma_n = F_n^{-1} - F^{-1}\), with \(F_n\) and \(F_n^{-1}\) being the empirical distribution and quantile functions, respectively, based on a sample of size \(n\) from \(F\). In the present paper, the process \(V_n\) is called the (generalized) Vervaat process, and its properties and applications in statistics are surveyed, with strong emphasis on asymptotics. Among the applications are the Lorenz curve, the Bonferroni curve, the total time on test curve and the Goldie curve. For the entire collection see [Zbl 0901.00049].