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: A refined continuity correction for the negative binomial distribution and asymptotics of the median

Abstract: In this paper, we prove a local limit theorem and a refined continuity correction for the negative binomial distribution. We present two applications of the results. First, we find the asymptotics of the median for a Negative Binomial (r,p) random variable jittered by a Uniform (0,1), which answers a problem left open in Coeurjolly \& Tr\'epanier (2020). This is used to construct a simple, robust and consistent estimator of the parameter $p$, when $r > 0$ is known. Second, we find an upper bound on the Le Cam distance between negative binomial and normal experiments.
Comments: 18 pages, 3 figures
Subjects: Statistics Theory (math.ST); Probability (math.PR)
MSC classes: 62E20, 62F12, 62F35, 62E15, 60F15, 62B15
Cite as: arXiv:2103.08846 [math.ST]
  (or arXiv:2103.08846v1 [math.ST] for this version)

Submission history

From: Frédéric Ouimet [view email]
[v1] Tue, 16 Mar 2021 04:25:50 GMT (325kb,D)

Link back to: arXiv, form interface, contact.