A Logodds Criterion for Selection of Diagnostic Tests

  • Christopher J. Whitaker
  • Ludmila I. Kuncheva
  • Peter D. Cockcroft
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3138)

Abstract

We propose a criterion for selection of independent binary diagnostic tests (signs). The criterion maximises the difference between the logodds for having the disease and the logodds for not having the disease. A parallel is drawn between the logodds criterion and the standard minimum error criterion. The error criterion is “progression non-monotone” which means that even for independent binary signs, the best set of two signs might not contain the single best sign. The logodds criterion is progression monotone, therefore the selection procedure consists of simply selecting the individually best features. A data set for scrapie in sheep is used as an illustration.

Keywords

feature selection combining diagnostic tests independent binary features logodds criterion veterinary medicine diagnosis ofscrapie in sheep 

References

  1. 1.
    Aha, D.W., Bankert, R.L.: A comparative evaluation of sequential feature selection algorithms. In: Proc. 5th International Workshop on AI and Statistics, Ft Lauderdale, FL, pp. 1–7 (1995)Google Scholar
  2. 2.
    Jain, A.K., Zongker, D.: Feature selection: evaluation, application and small sample performance. IEEE Transactions on PAMI 19(2), 153–158 (1997)Google Scholar
  3. 3.
    Kuncheva, L.I., Cockcroft, P.D., Whitaker, C.J., Hoare, Z.S.: Pre-selection of independent binary features in differential diagnosis (submitted)Google Scholar
  4. 4.
    In: Roli, F., Kittler, J. (eds.) Proc. 2nd International Workshop on Multiple Classifier Systems (MCS 2001). Lecture Notes in Computer Science LNCS, vol. 2096, Springer, Cambridge (2001)Google Scholar
  5. 5.
    Toussaint, G.T.: Note on optimal selection of independent binary-valued features for pattern recognition. IEEE Transactions on Information Theory 17, 618 (1971)Google Scholar
  6. 6.
    White, M.E.: Diagnosis, information management, teaching and record coding using the consultant database. Canadian Veterinary Journal 29, 271–273 (1988)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Christopher J. Whitaker
    • 1
  • Ludmila I. Kuncheva
    • 1
  • Peter D. Cockcroft
    • 2
  1. 1.School of InformaticsUniversity of WalesBangorUK
  2. 2.Department of Clinical Veterinary MedicineUniversity of CambridgeUK

Personalised recommendations