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: Parameter recovery in two-component contamination mixtures: the $\mathbb{L}^2$ strategy

Abstract: In this paper, we consider a parametric density contamination model. We work with a sample of i.i.d. data with a common density, $f^\star =(1-\lambda^\star) \phi + \lambda^\star \phi(.-\mu^\star)$, where the shape $\phi$ is assumed to be known. We establish the optimal rates of convergence for the estimation of the mixture parameters $(\lambda^\star,\mu^\star)$. In particular, we prove that the classical parametric rate $1/\sqrt{n}$ cannot be reached when at least one of these parameters is allowed to tend to $0$ with $n$.
Comments: 48 pages, 4 figures
Subjects: Statistics Theory (math.ST)
Cite as: arXiv:1604.00306 [math.ST]
  (or arXiv:1604.00306v2 [math.ST] for this version)

Submission history

From: Sebastien Gadat [view email]
[v1] Fri, 1 Apr 2016 16:03:11 GMT (1031kb,D)
[v2] Thu, 22 Feb 2018 20:37:13 GMT (1048kb,D)
[v3] Wed, 21 Nov 2018 14:30:44 GMT (1049kb,D)

Link back to: arXiv, form interface, contact.