skip to main content
10.1145/1276958.1277334acmconferencesArticle/Chapter ViewAbstractPublication PagesgeccoConference Proceedingsconference-collections
Article

MDL-based fitness for feature construction

Published: 07 July 2007 Publication History

Abstract

Primitive data representation of real-world data facilitates attribute interactions, which make information opaque to most learners. Feature Construction (FC) aims to abstract and encapsulate interactions into new features and outline them to the learner. When a GA is applied to perform FC, the goal is to generate features that facilitate more accurate learning. Then the GA's fitness function should estimate the quality of the constructed features. We propose a new fitness function based on Minimum Description Length (MDL). This fitness is incorporated in MFE2/GA to improve its accuracy. The new system is compared with other systems based on Entropy or error-rate fitness.

References

[1]
A. A. Freitas. Understanding the crucial role of attribute interaction in data mining. AI Review, 16(3):177--199, Nov. 2001.
[2]
R. J. Quinlan. C4.5: Programs for Machine Learning. Morgan Kaufmann, San Mateo, California, 1993.
[3]
J. Rissanen. A universal prior for integers and estimation by minimum description length. The Annals of Statistics, 11(2):416--431, Jun. 1983.
[4]
L. S. Shafti and E. Pérez. Reducing complex attribute interaction through non-algebraic feature construction. In Proc. of the IASTED--AIA, pages 359--365, Innsbruck, Austria, Feb. 2007. Acta Press.
[5]
B. Zupan, M. Bohanec, I. Bratko, and J. Demsar. Learning by discovering concept hierarchies. Artificial Intelligence, 109(1--2):211--242, 1999.

Recommendations

Comments

Information & Contributors

Information

Published In

cover image ACM Conferences
GECCO '07: Proceedings of the 9th annual conference on Genetic and evolutionary computation
July 2007
2313 pages
ISBN:9781595936974
DOI:10.1145/1276958

Sponsors

Publisher

Association for Computing Machinery

New York, NY, United States

Publication History

Published: 07 July 2007

Permissions

Request permissions for this article.

Check for updates

Author Tags

  1. Entropy
  2. MDL
  3. attribute interaction
  4. feature construction
  5. fitness function
  6. machine learning

Qualifiers

  • Article

Conference

GECCO07
Sponsor:

Acceptance Rates

GECCO '07 Paper Acceptance Rate 266 of 577 submissions, 46%;
Overall Acceptance Rate 1,669 of 4,410 submissions, 38%

Contributors

Other Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

  • 0
    Total Citations
  • 125
    Total Downloads
  • Downloads (Last 12 months)0
  • Downloads (Last 6 weeks)0
Reflects downloads up to 19 Feb 2025

Other Metrics

Citations

View Options

Login options

View options

PDF

View or Download as a PDF file.

PDF

eReader

View online with eReader.

eReader

Figures

Tables

Media

Share

Share

Share this Publication link

Share on social media