Authors: Bissantz, Nicolai¹; Dümbgen, Lutz²; Holzmann, Hajo³; Munk, Axel³

Source: Journal of the Royal Statistical Society: Series B (Statistical Methodology), Volume 69, Number 3, June 2007 , pp. 483-506(24)

Publisher: Wiley-Blackwell

Abstract:

Summary. 

Uniform confidence bands for densities f via non-parametric kernel estimates were first constructed by Bickel and Rosenblatt. In this paper this is extended to confidence bands in the deconvolution problem g=f*ψ for an ordinary smooth error density ψ. Under certain regularity conditions, we obtain asymptotic uniform confidence bands based on the asymptotic distribution of the maximal deviation (L_∞-distance) between a deconvolution kernel estimator and f. Further consistency of the simple non-parametric bootstrap is proved. For our theoretical developments the bias is simply corrected by choosing an undersmoothing bandwidth. For practical purposes we propose a new data-driven bandwidth selector that is based on heuristic arguments, which aims at minimizing the L_∞-distance between and f. Although not constructed explicitly to undersmooth the estimator, a simulation study reveals that the bandwidth selector suggested performs well in finite samples, in terms of both area and coverage probability of the resulting confidence bands. Finally the methodology is applied to measurements of the metallicity of local F and G dwarf stars. Our results confirm the `G dwarf problem', i.e. the lack of metal poor G dwarfs relative to predictions from `closed box models' of stellar formation.

Keywords: Astrophysics; Asymptotic normality; Bandwidth selection; Bootstrap; Confidence band; Deconvolution; Non-parametric density estimation

Document Type: Research article

DOI: http://dx.doi.org/10.1111/j.1467-9868.2007.599.x

Affiliations: 1: Georg-August-Universität, Göttingen, and Ruhr-Universität Bochum, Germany 2: University of Bern, Switzerland 3: Georg-August-Universität, Göttingen, Germany

Publication date: 2007-06-01

Related content

• In this: publication
• By this: publisher
• In this Subject: Mathematics and Statistics
• By this author: Bissantz, Nicolai ;  Dümbgen, Lutz ;  Holzmann, Hajo ;  Munk, Axel

Website © Publishing Technology. Article copyright remains with the publisher, society or author(s) as specified within the article.

• Terms and conditions
• Privacy policy
• Information for advertisers