On Stopping Rules in Dependency-Aware Feature Ranking

0399659 - ÚTIA 2014 RIV DE eng C - Conference Paper (international conference)
Somol, Petr - Grim, Jiří - Filip, Jiří - Pudil, P.
On Stopping Rules in Dependency-Aware Feature Ranking.
Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications. Heidelberg: Springer, 2013, s. 286-293. Lecture Notes in Computer Science, 8258. ISBN 978-3-642-41821-1.
[CIARP 2013, Iberoamerican Congress on Pattern Recognition /18./. Havana (CU), 20.11.2013-23.11.2013]
R&D Projects: GA ČR GAP103/11/0335
Institutional support: RVO:67985556
Keywords : dimensionality reduction * feature selection * randomization and stopping rule
Subject RIV: BD - Theory of Information
http://library.utia.cas.cz/separaty/2013/RO/somol-0399659.pdf

Feature Selection in very-high-dimensional or small sample problems is particularly prone to computational and robustness complications. It is common to resort to feature ranking approaches only or to randomization techniques. A recent novel approach to the randomization idea in form of Dependency-Aware Feature Ranking (DAF) has shown great potential in tackling these problems well. Its original definition, however, leaves several technical questions open. In this paper we address one of these questions: how to define stopping rules of the randomized computation that stands at the core of the DAF method. We define stopping rules that are easier to interpret and show that the number of randomly generated probes does not need to be extensive.
Permanent Link: http://hdl.handle.net/11104/0226952