Asymptotic tail behavior of uniform multivariate empirical processes (Q2639402): Difference between revisions

Let \(\alpha_ n\) be the empirical process of independent uniformly distributed random vectors on the unit square \(I^ 2\). The authors study the asymptotic distribution of the random variable \[ \sup | \alpha_ n(s,t)| /(s^{\nu}t^{\mu}L(s)G(s)), \] when the sup is taken over various subintervals of \(I^ 2\). They show that the limit is given in terms of a two time parameter Wiener process if \(\nu\) and \(\mu\) are smaller than 1/2 and is determined by a Poisson process if both \(\nu\) and \(\mu\) are greater than 1/2.