Some Novel Limiting Distributions Arising in Order Restricted Inference

Sayan Ghosh, Ori Davidov

Research output: Contribution to journalArticlepeer-review

Abstract

In this paper, we develop a methodology for testing the hypothesis that the true value of a parameter lies in the union of multiple cones against the alternative that it does not. We propose a test statistic for such problems and derive its novel asymptotic null distribution. The least favourable asymptotic null value and the corresponding least favourable asymptotic null distribution are obtained. The proposed test is uniformly more powerful than conventional tests discussed in the literature. Some illustrative examples are provided and a simulation study evaluating its performance is presented.

Original languageEnglish
Pages (from-to)45-66
Number of pages22
JournalStatistics and Applications
Volume22
Issue number3
StatePublished - 2024

Bibliographical note

Publisher Copyright:
© 2024, Society of Statistics, Computer and Applications. All rights reserved.

Keywords

  • Asymptotic Distribution
  • Convex Cones
  • Hypothesis Testing
  • Least Favourable Configuration

ASJC Scopus subject areas

  • Statistics and Probability

Fingerprint

Dive into the research topics of 'Some Novel Limiting Distributions Arising in Order Restricted Inference'. Together they form a unique fingerprint.

Cite this