Abstract <p>This study aims to investigate the convergence properties of the exact closed-form formula for the distribution of indices in a sorted sample being medians of Poisson-bootstrapped subsamples toward the binomial distribution. We find that although the two distributions do not coincide exactly, their asymptotic behaviors are of the same order, specifically <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(O\left(1/{\sqrt{n}}\right)\)</EquationSource> <!--LobJMat2561350Pile-m1--> </InlineEquation> as the sample size <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(n\)</EquationSource> <!--LobJMat2561350Pile-m2--> </InlineEquation> increases. The findings have significant implications for statistical inference and resampling techniques, as they provide deeper insights into the efficiency and accuracy of the Poisson bootstrap method.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

On Convergence of Different Bootstrap Approximations for Medians

  • I. E. Pile

摘要

Abstract

This study aims to investigate the convergence properties of the exact closed-form formula for the distribution of indices in a sorted sample being medians of Poisson-bootstrapped subsamples toward the binomial distribution. We find that although the two distributions do not coincide exactly, their asymptotic behaviors are of the same order, specifically \(O\left(1/{\sqrt{n}}\right)\) as the sample size \(n\) increases. The findings have significant implications for statistical inference and resampling techniques, as they provide deeper insights into the efficiency and accuracy of the Poisson bootstrap method.