A note on min--maxbias estimators in approximately linear models
Gilbert Bassett
Statistics & Probability Letters, 1994, vol. 21, issue 1, 27-28
Abstract:
The approximately linear model represents deviations from the ideal linear model by a vector contained in a prescribed bias-ball. In a recent paper Mathew and Nordstrom (1993) proposed min--maxbias estimators in which a criterion function is defined by maximizing errors over the bias-ball. When the Chebyshev norm defines the bias-ball they found the least absolute deviation or L1 estimator to be identical to its maxbias version. This was thought to be a robustness property since it contrasts with least squares where the maxbias criterion is a combination of L1 and the sum of squares. In this paper it is shown, however, that equivalence between the L1 estimator and its min--maxbias version is not special to L1 and that the equivalence is valid for estimates that are not robust. Hence, while the L1 estimate does have desirable robustness properties the equivalence to its min--maxbias version cannot be counted as one of them.
Keywords: Approximately; linear; model; Least; squares; Least; absolute; deviation (search for similar items in EconPapers)
Date: 1994
References: Add references at CitEc
Citations:
Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/0167-7152(94)90054-X
Full text for ScienceDirect subscribers only
Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
HTML/Text
Persistent link: https://EconPapers.repec.org/RePEc:eee:stapro:v:21:y:1994:i:1:p:27-28
Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/supportfaq.cws_home/regional
https://shop.elsevie ... _01_ooc_1&version=01
Access Statistics for this article
Statistics & Probability Letters is currently edited by Somnath Datta and Hira L. Koul
More articles in Statistics & Probability Letters from Elsevier
Bibliographic data for series maintained by Catherine Liu ().