Scholarlink

<p>We propose a simple and distribution-independent method for identifying elite elements in a ranked population. Inspired by the h-index and based on the intersection of normalized value and relative rank curves, the method yields an elite threshold <i>η</i>, which designates the top <i>h</i> = <i>η·n</i> elements of the sample. We demonstrate its theoretical behavior in three canonical distributions (uniform, exponential, Pareto) and illustrate its application to real-world data from the SCImago journal database. A refined estimation based on the Hill tail-index provides a robust alternative to the empirical <i>η</i>, and the resulting indicator is shown to correlate well with traditional concentration measures such as the Gini index. The method offers a conceptually clear and computationally simple approach to elite selection and concentration measurement in heavy-tailed distributions.</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

A possible definition of the elite in a ranked population

  • 期刊论文

  • 发表日期：

    2026年1月14日

查看全文
  • András Schubert

摘要

We propose a simple and distribution-independent method for identifying elite elements in a ranked population. Inspired by the h-index and based on the intersection of normalized value and relative rank curves, the method yields an elite threshold η, which designates the top h = η·n elements of the sample. We demonstrate its theoretical behavior in three canonical distributions (uniform, exponential, Pareto) and illustrate its application to real-world data from the SCImago journal database. A refined estimation based on the Hill tail-index provides a robust alternative to the empirical η, and the resulting indicator is shown to correlate well with traditional concentration measures such as the Gini index. The method offers a conceptually clear and computationally simple approach to elite selection and concentration measurement in heavy-tailed distributions.