scientific article; zbMATH DE number 1547070

From MaRDI portal
Publication:4522492





Let \(X_1,X_2,\dots, X_n\) be \(n\) i.i.d. (independent, identically distributed) random variables having the common distribution function \(F(x)={\mathbf P}(X_1\leq x)\), and let \(X_{(1)}\leq X_{(2)}\leq\cdots\leq X_{(n)}\) be the corresponding order statistics. A well-known result of B. V. Gnedenko (1943) says that there exist two normalization real constants \(a_n> 0\) and \(b_n\) such that, when \(n\to\infty\), \((X_{(n)}- b_n)/a_n\) has as limit distribution a function \(G\) given by: \[ {\mathbf P}(X_{(n)}\leq a_nx+ b_n)= F^n(a_nx+ b_n)\to G(x).\tag{\(*\)} \] The aim of the present paper is to evaluate the convergence rate of the limit process \((*)\), using a Poisson approximation of the binomial distribution. Another result of the paper refers to the convergence rate for the limit joint distribution of the last \(k\)-order statistics, in the case of i.i.d. random variables. This represents a generalization of a known result for two Poisson random variables (D. Freedman; 1974).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4522492)