We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.ST

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Statistics Theory

Title: Asymptotic coverage probabilities of bootstrap percentile confidence intervals for constrained parameters

Abstract: The asymptotic behaviour of the commonly used bootstrap percentile confidence interval is investigated when the parameters are subject to linear inequality constraints. We concentrate on the important one- and two-sample problems with data generated from general parametric distributions in the natural exponential family. The focus of this paper is on quantifying the coverage probabilities of the parametric bootstrap percentile confidence intervals, in particular their limiting behaviour near boundaries. We propose a local asymptotic framework to study this subtle coverage behaviour. Under this framework, we discover that when the true parameters are on, or close to, the restriction boundary, the asymptotic coverage probabilities can always exceed the nominal level in the one-sample case; however, they can be, remarkably, both under and over the nominal level in the two-sample case. Using illustrative examples, we show that the results provide theoretical justification and guidance on applying the bootstrap percentile method to constrained inference problems.
Comments: 22 pages, 6 figures
Subjects: Statistics Theory (math.ST); Computation (stat.CO); Methodology (stat.ME)
Cite as: arXiv:1712.02469 [math.ST]
  (or arXiv:1712.02469v1 [math.ST] for this version)

Submission history

From: Chunlin Wang [view email]
[v1] Thu, 7 Dec 2017 01:51:35 GMT (132kb)

Link back to: arXiv, form interface, contact.