Please use this identifier to cite or link to this item: https://doi.org/10.1093/biomet/80.4.869
DC FieldValue
dc.titleOn median-type estimators of direction for the von mises-fisher distribution
dc.contributor.authorChan, Y.M.
dc.contributor.authorHe, X.
dc.date.accessioned2014-10-28T02:40:35Z
dc.date.available2014-10-28T02:40:35Z
dc.date.issued1993-12
dc.identifier.citationChan, Y.M., He, X. (1993-12). On median-type estimators of direction for the von mises-fisher distribution. Biometrika 80 (4) : 869-875. ScholarBank@NUS Repository. https://doi.org/10.1093/biomet/80.4.869
dc.identifier.issn00063444
dc.identifier.urihttp://scholarbank.nus.edu.sg/handle/10635/103723
dc.description.abstractSUMMARY: We study three median-type estimators of direction: the normalized spatial median, L1-estimator on the sphere and spherical median. Their asymptotic distributions, efficiencies, influence functions and sensitivities are given at the von Mises-Fisher model. A simulation study is also carried out to compare their finite sample performance in terms of mean squared errors. We conclude that overall the normalized spatial median is the preferred estimator. © 1993 Biometrika Trust.
dc.sourceScopus
dc.subjectAsymptotic distribution
dc.subjectDirectional data
dc.subjectEfficiency
dc.subjectGross-error sensitivity
dc.subjectInfluence function
dc.subjectL1-estimator
dc.subjectSpatial median
dc.subjectSpherical median
dc.subjectVon Mises-Fisher distribution
dc.typeArticle
dc.contributor.departmentMATHEMATICS
dc.description.doi10.1093/biomet/80.4.869
dc.description.sourcetitleBiometrika
dc.description.volume80
dc.description.issue4
dc.description.page869-875
dc.description.codenBIOKA
dc.identifier.isiutA1993NA47300015
Appears in Collections:Staff Publications

Show simple item record
Files in This Item:
There are no files associated with this item.

SCOPUSTM   
Citations

14
checked on Mar 17, 2023

WEB OF SCIENCETM
Citations

11
checked on Mar 17, 2023

Page view(s)

207
checked on Mar 16, 2023

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.