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: Towards Practical Mean Bounds for Small Samples

Abstract: Historically, to bound the mean for small sample sizes, practitioners have had to choose between using methods with unrealistic assumptions about the unknown distribution (e.g., Gaussianity) and methods like Hoeffding's inequality that use weaker assumptions but produce much looser (wider) intervals. In 1969, Anderson (1969) proposed a mean confidence interval strictly better than or equal to Hoeffding's whose only assumption is that the distribution's support is contained in an interval $[a,b]$. For the first time since then, we present a new family of bounds that compares favorably to Anderson's. We prove that each bound in the family has {\em guaranteed coverage}, i.e., it holds with probability at least $1-\alpha$ for all distributions on an interval $[a,b]$. Furthermore, one of the bounds is tighter than or equal to Anderson's for all samples. In simulations, we show that for many distributions, the gain over Anderson's bound is substantial.
Comments: This is an extended work of our ICML 2021 paper "Towards Practical Mean Bounds for Small Samples"
Subjects: Statistics Theory (math.ST)
Cite as: arXiv:2106.03163 [math.ST]
  (or arXiv:2106.03163v5 [math.ST] for this version)

Submission history

From: My Phan [view email]
[v1] Sun, 6 Jun 2021 16:07:20 GMT (4561kb,D)
[v2] Sun, 13 Jun 2021 14:38:34 GMT (4498kb,D)
[v3] Sun, 1 Aug 2021 16:36:40 GMT (4498kb,D)
[v4] Wed, 22 Sep 2021 20:45:49 GMT (4498kb,D)
[v5] Mon, 25 Oct 2021 19:54:14 GMT (4498kb,D)

Link back to: arXiv, form interface, contact.