Maximum likelihood estimation of the linear model with equicorrelated errors

  • Giuseppe De Luca
  • , Jan R. Magnus
  • , Andrey L. Vasnev*
  • *Corresponding author for this work

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

Abstract.: We provide a simple proof that the maximum-likelihood estimator in the linear model with equicorrelated errors does not exist, and explain why not.

Original languageEnglish
Pages (from-to)6295-6302
Number of pages8
JournalCommunications in Statistics - Theory and Methods
Volume54
Issue number19
DOIs
Publication statusPublished - 2025

Bibliographical note

Publisher Copyright:
© 2025 The Author(s). Published with license by Taylor & Francis Group, LLC.

Funding

Giuseppe De Luca acknowledges financial support from MUIR, Italy, PRIN PRJ-1446 and PRIN PRJ-1547 at the University of Palermo. We thank the editor and reviewer for their helpful comments and suggestions.

Funders
MUIR

    Keywords

    • compound symmetry
    • Equicorrelation
    • maximum likelihood
    • non existence
    • variance components

    Fingerprint

    Dive into the research topics of 'Maximum likelihood estimation of the linear model with equicorrelated errors'. Together they form a unique fingerprint.

    Cite this