Show simple item record

dc.contributor.authorEstivill-Castro, V
dc.contributor.authorYang, J
dc.contributor.editorHeikki Mannila
dc.date.accessioned2017-05-03T14:15:42Z
dc.date.available2017-05-03T14:15:42Z
dc.date.issued2004
dc.date.modified2010-08-09T07:17:25Z
dc.identifier.issn1384-5810
dc.identifier.doi10.1023/B:DAMI.0000015869.08323.b3
dc.identifier.urihttp://hdl.handle.net/10072/5182
dc.description.abstractGeneral purpose and highly applicable clustering methods are usually required during the early stages of knowledge discovery exercises. k-MEANS has been adopted as the prototype of iterative model-based clustering because of its speed, simplicity and capability to work within the format of very large databases. However, k-MEANS has several disadvantages derived from its statistical simplicity. We propose an algorithm that remains very efficient, generally applicable, multidimensional but is more robust to noise and outliers. We achieve this by using medians rather than means as estimators for the centers of clusters. Comparison with k-MEANS, EXPECTATION and MAXIMIZATION sampling demonstrates the advantages of our algorithm.
dc.description.peerreviewedYes
dc.description.publicationstatusYes
dc.format.extent905036 bytes
dc.format.extent61987 bytes
dc.format.mimetypeapplication/pdf
dc.format.mimetypetext/plain
dc.languageEnglish
dc.language.isoeng
dc.publisherSpringer
dc.publisher.placeDORDRECHT, NETHERLAN
dc.relation.ispartofpagefrom127
dc.relation.ispartofpageto150
dc.relation.ispartofissue2
dc.relation.ispartofjournalData Mining and Knowledge Discovery
dc.relation.ispartofvolume8
dc.subject.fieldofresearchData management and data science
dc.subject.fieldofresearchInformation systems
dc.subject.fieldofresearchcode4605
dc.subject.fieldofresearchcode4609
dc.titleFast and Robust General Purpose Clustering Algorithms
dc.typeJournal article
dc.type.descriptionC1 - Articles
dc.type.codeC - Journal Articles
gro.facultyGriffith Sciences, School of Information and Communication Technology
gro.rights.copyright© Springer 2004. This is the author-manuscript version of this paper. The original publication is available at www.springerlink.com
gro.date.issued2004
gro.hasfulltextFull Text
gro.griffith.authorEstivill-Castro, Vladimir


Files in this item

This item appears in the following Collection(s)

  • Journal articles
    Contains articles published by Griffith authors in scholarly journals.

Show simple item record