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


Current browse context:


Change to browse by:

References & Citations


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

Mathematics > Statistics Theory

Title: Robust mixture regression based on the skew t distribution

Abstract: In this study, we propose a robust mixture regression procedure based on the skew t distribution to model heavy-tailed and/or skewed errors in a mixture regression setting. Using the scale mixture representation of the skew t distribution, we give an Expectation Maximization (EM) algorithm to compute the maximum likelihood (ML) estimates for the paramaters of interest. The performance of proposed estimators is demonstrated by a simulation study and a real data example.
Comments: 15
Subjects: Statistics Theory (math.ST); Methodology (stat.ME)
DOI: 10.15446/rce.v40n1.53580
Cite as: arXiv:1512.00377 [math.ST]
  (or arXiv:1512.00377v1 [math.ST] for this version)

Submission history

From: Fatma Zehra Doğru [view email]
[v1] Tue, 1 Dec 2015 18:34:00 GMT (549kb)

Link back to: arXiv, form interface, contact.