The Instance Easiness of Supervised Learning for Cluster Validity

File Size Format
70643_1.pdf 646Kb Adobe PDF View
Title The Instance Easiness of Supervised Learning for Cluster Validity
Author Estivill-Castro, Vladimir
Journal Name Lecture Notes in Computer Science
Editor Philippe Lenca
Year Published 2012
Place of publication Germany
Publisher Springer
Abstract “The statistical problem of testing cluster validity is essentially unsolved” [5]. We translate the issue of gaining credibility on the output of un-supervised learning algorithms to the supervised learning case. We introduce a notion of instance easiness to supervised learning and link the validity of a clustering to how its output constitutes an easy instance for supervised learning. Our notion of instance easiness for supervised learning extends the notion of stability to perturbations (used earlier for measuring clusterability in the un-supervised setting). We follow the axiomatic and generic formulations for cluster-quality measures. As a result, we inform the trust we can place in a clustering result using standard validity methods for supervised learning, like cross validation.
Peer Reviewed Yes
Published Yes
Alternative URI http://dx.doi.org/10.1007/978-3-642-28320-8_17
Copyright Statement Copyright 2011 Springer Berlin / Heidelberg. This is the author-manuscript version of this paper. Reproduced in accordance with the copyright policy of the publisher. The original publication is available at www.springerlink.com
Volume 7104
Page from 197
Page to 208
ISSN 0302-9743
Date Accessioned 2012-03-05
Date Available 2013-01-15T00:36:11Z
Language en_US
Research Centre Institute for Integrated and Intelligent Systems
Faculty Faculty of Science, Environment, Engineering and Technology
Subject Pattern Recognition and Data Mining
URI http://hdl.handle.net/10072/45878
Publication Type Journal Articles (Refereed Article)
Publication Type Code c1

Brief Record

Griffith University copyright notice