Research output

Conference Paper (in Formal Publication) ()

A Targeted Estimation of Distribution Algorithm Compared to Traditional Methods in Feature Selection

Citation
Neumann G & Cairns D (2016) A Targeted Estimation of Distribution Algorithm Compared to Traditional Methods in Feature Selection In: Madani K, Dourado A, Rosa A, Filipe J, Kacprzyk J (ed.) Computational Intelligence: Revised and Selected Papers of the International Joint Conference, IJCCI 2013, Vilamoura, Portugal, September 20-22, 2013, Cham, Switzerland: Springer. 5th International Joint Conference on Computational Intellegience, IJCCI 2013, 20.9.2013 - 22.9.2013, Vilamoura, Portugal, pp. 83-103.

Abstract
The Targeted Estimation of Distribution Algorithm (TEDA) introduces into an EDA/GA hybrid framework a ‘Targeting’ process, whereby the number of active genes, or ‘control points’, in a solution is driven in an optimal direction. For larger feature selection problems with over a thousand features, traditional methods such as forward and backward selection are inefficient. Traditional EAs may perform better but are slow to optimize if a problem is sufficiently noisy that most large solutions are equally ineffective and it is only when much smaller solutions are discovered that effective optimization may begin. By using targeting, TEDA is able to drive down the feature set size quickly and so speeds up this process. This approach was tested on feature selection problems with between 500 and 20,000 features using all of these approaches and it was confirmed that TEDA finds effective solutions significantly faster than the other approaches.

Keywords
Estimation of distribution algorithms; Feature selection; Evolutionary computation; Genetic algorithms; Hybrid algorithms

StatusPublished
EditorMadani K, Dourado A, Rosa A, Filipe J, Kacprzyk J
AuthorsNeumann Geoffrey, Cairns David
Title of seriesStudies in Computational Intelligence
Number in series613
Publication date2016
Date of public distribution09/2013
URLhttp://link.springer.com/…-3-319-23392-5_5
PublisherSpringer
Place of publicationCham, Switzerland
ISSN of series 1860-949X
ISBN 978-3-319-23391-8
eISBN 978-3-319-23392-5
LanguageEnglish
© University of Stirling FK9 4LA Scotland UK • Telephone +44 1786 473171 • Scottish Charity No SC011159
My Portal