Advertisement

Evolving Receiver Operating Characteristics for Data Fusion

  • William B. Langdon
  • Bernard F. Buxton
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2038)

Abstract

It has been suggested that the “Maximum Realisable Receiver Operating Characteristics” for a combination of classifiers is the convex hull of their individual ROCs [Scott et al., 1998]. As expected in at least some cases better ROCs can be produced. We show genetic programming (GP) can automatically produce a combination of classifiers whose ROC is better than the convex hull of the supplied classifier’s ROCs.

Keywords

False Alarm Convex Hull False Positive Rate Receiver Operating Characteristic Receiver Operating Characteristic Curve 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Burbidge et al., 2001._R. Burbidge, M. Trotter, B. Buxton, and S. Holden. Drug design by machine learning: Support vector machines for pharmaceutical data analysis. Computers in Chemistry, 2001. Under review.Google Scholar
  2. Freitas, 1999.
    Alex A. Freitas. Data mining with evolutionary algorithms: Research directions. Technical Report WS-99-06, AAAI, Orlando, USA, July 1999. Papers from the AAAI Workshop http://www.aaai.org/Press/Reports/Workshops/ws-99-06.html.
  3. Langdon and Buxton, 2001.
    W.B. Langdon and B.F. Buxton. Genetic programming for combining classifiers. 2001. Submitted.Google Scholar
  4. Langdon et al., 1999._William B. Langdon, Terry Soule, Riccardo Poli, and James A. Foster. The evolution of size and shape. In Lee Spector, William B. Langdon, UnaMay O’Reilly, and Peter J. Angeline, editors, Advances in Genetic Programming 3, chapter 8, pages 163–190. MIT Press, Cambridge, MA, USA, June 1999.Google Scholar
  5. Langdon, 1998.
    William B. Langdon. Data Structures and Genetic Programming: Genetic Programming + Data Structures = Automatic Programming!, volume 1 of Genetic Programming. Kluwer, Boston, 1998.Google Scholar
  6. Langdon, 2000a.
    W.B. Langdon. Problems with ROC convex hulls. http://www.cs.ucl.ac.uk/staff/W.Langdon/roc.ps.gz, 24 May 2000.
  7. Langdon, 2000b.
    William B. Langdon. Size fair and homologous tree genetic programming crossovers. Genetic Programming And Evolvable Machines, 1(1/2):95–119, April 2000.zbMATHCrossRefGoogle Scholar
  8. Scott et al., 1998._M.J.J. Scott, M. Niranjan, and R.W. Prager. Realisable classifiers: Improving operating performance on variable cost problems. In Paul H. Lewis and Mark S. Nixon, editors, Proceedings of the Ninth British Machine Vision Conference, volume 1, pages 304–315, University of Southampton, UK, 14-17 September 1998.Google Scholar
  9. Swets et al., 2000._John A. Swets, Robyn M. Dawes, and John Monahan. Better decisions through science. Scientific American, pages 70–75, October 2000.Google Scholar
  10. Yusoff et al., 1998._Y. Yusoff, J. Kittler, and W. Christmas. Combining multiple experts for classifying shot changes in video sequences. In IEEE International Conference on Multimedia Computing and Systems, volume II, Florence, Italy, 7-11 June 1998.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2001

Authors and Affiliations

  • William B. Langdon
    • 1
  • Bernard F. Buxton
    • 1
  1. 1.Computer Science, University CollegeLondonUK

Personalised recommendations