Improved confidence estimators for confidence sets of location parameters

Hsiuying Wang*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Consider a p-dimensional location family symmetrical about θ. Let Ct(X) be a 1 - α confidence set {θ: t(X) - θ ≤c} of θ, where t(X) is some reasonable estimator of θ. Traditionally, the confidence coefficient 1 - α, which is data independent, is used to be the report for the confidence of Ct(X). In this paper, some improved confidence reports are provided for p ≥ 5. These results are related to Robinson (Ann. Statistic 7 (1979) 756). The normal case discussed in Robinson (Ann. Statistic 7 (1979) 756) is a special case of the results of this paper. Moreover, some admissibility results when p≤4 are also present in this paper.

Original languageEnglish
Pages (from-to)95-107
Number of pages13
JournalJournal of Statistical Planning and Inference
Volume128
Issue number1
DOIs
StatePublished - 15 Jan 2005

Keywords

  • Admissibility
  • Confidence coefficient
  • t-distribution

Fingerprint

Dive into the research topics of 'Improved confidence estimators for confidence sets of location parameters'. Together they form a unique fingerprint.

Cite this