On robust cross-validation for nonparametric smoothing

Simple item page
dc.contributor.authorFried, Roland
dc.contributor.authorMorell, Oliver
dc.contributor.authorOtto, Dennis
dc.date.accessioned2010-05-12T07:16:49Z
dc.date.available2010-05-12T07:16:49Z
dc.date.issued2010-05-12T07:16:49Z
dc.description.abstractProcedures for local-constant smoothing are investigated in a broad variety of data situations with outliers and jumps. Moving window and nearest neighbour versions of mean and median smoothers are considered, as well as double window and linear hybrid smoothers. For the choice of the window width or the number of neighbours the different estimators are combined with each of several cross-validation criteria like least squares, least absolute deviations, and median-cross-validation. It is identified, which method works best in which data scenarios. Although there is not a single overall best robust smoothing procedure, a robust cross-validation criterion, called least trimmed squares-cross-validation, gives good results for most smoothing methods and data situations, with cross-validation based on least absolute deviations being a strong competitor, particularly if there are jumps, but little problems with outliers in the data.en
dc.identifier.urihttp://hdl.handle.net/2003/27168
dc.identifier.urihttp://dx.doi.org/10.17877/DE290R-8729
dc.language.isoenen
dc.relation.ispartofseriesDiscussion Paper / SFB 823;17/2010
dc.subjectJumpen
dc.subjectLeast trimmed squares-cross-validationen
dc.subjectLocal-constant smoothingen
dc.subjectLTS-CVen
dc.subjectOutlieren
dc.subjectRobust smoothing procedureen
dc.subject.ddc310
dc.subject.ddc330
dc.subject.ddc620
dc.titleOn robust cross-validation for nonparametric smoothingen
dc.typeTextde
dc.type.publicationtypereportde
dcterms.accessRightsopen access

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
DP_1710_SFB823_morell_otto_fried.pdf
Size:
387.5 KB
Format:
Adobe Portable Document Format
Description:
DNB
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.09 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

Sonderforschungsbereich (SFB) 823