Skip to main navigation Skip to search Skip to main content

Quantile-based empirical likelihood for ROC curves

  • Pengfei LIU
  • , Yiming LIU*
  • , Zhi LIU
  • , Yujie LIANG
  • *Corresponding author for this work

Research output: Journal Publications › Journal Article (refereed) › peer-review

Abstract

We develop pointwise inference for sensitivity at a fixed false positive rate by profiling the unknown cutoff in an existing two-sample quantile empirical likelihood (QEL) framework. For a smoothed criterion, we establish the asymptotic distribution of the profiled cutoff and a chi-square limit with one degree of freedom under explicit bandwidth conditions. Interval calibration requires no separate plug-in variance estimate, and computation uses a one-dimensional cutoff search. In the examined designs, QEL has smaller mean absolute coverage errors than the compared influence-function and Bayesian alternatives, generally with wider intervals. A hepatitis C illustration is included; simulations, sensitivity checks and proofs are provided in the supplementary material.
Original languageEnglish
Article number110997
Number of pages5
JournalStatistics and Probability Letters
Volume242
Early online date24 Oct 2026
DOIs
Publication statusE-pub ahead of print - 24 Oct 2026
Externally publishedYes

Bibliographical note

Publisher Copyright:
© 2026 Elsevier B.V.

Funding

Yiming Liu, was partially supported by grants 2026A1515011522 of Natural Science Foundation of Guangdong Province, China and 12301340 of NSFC. Pengfei Liu was supported by the 12371266 and 52034007 of NSFC, the State Scholarship funded by the China Scholarship Council (CSC No. 202408320318), the Jiangsu Social Science Applied Research Boutique Project: Special Topic on Technological Innovation Services and New Quality Productive Forces (Grant No. 25SXC-039), and the Priority Academic Program Development of Jiangsu Higher Education Institutions. Zhi Liu was partially supported by MYRG-GRG2026-00029-FST and MYRG-GRG2025-00093-FST .

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

  1. SDG 3 - Good Health and Well-being
    SDG 3 Good Health and Well-being

Keywords

  • Empirical likelihood
  • Quantiles
  • ROC curves
  • Smoothing

Fingerprint

Dive into the research topics of 'Quantile-based empirical likelihood for ROC curves'. Together they form a unique fingerprint.

Cite this