Influence diagnostic for censored regression models

Chihwa Kao

Research output: Contribution to journalArticle

Abstract

This paper proposes a diagnostic measure to detect outliers either in the response variable or in the design space both for Schmee-Hahn estimator and Buckley-James estimator in censored regressions. The diagnostics consists of the leverage point and standardized predicted residual.

Original languageEnglish (US)
Pages (from-to)337-342
Number of pages6
JournalStatistics and Probability Letters
Volume3
Issue number6
DOIs
StatePublished - Oct 1985

    Fingerprint

Keywords

  • Buckley-James estimator
  • EM algorithm
  • Mahalanobis distance
  • Schmee-Hahn estimator
  • outliers
  • standardized predicted residual

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this