Exploiting context when learning to classify

  • Peter D. Turney
Position Papers Inductive Learning and Applications
Part of the Lecture Notes in Computer Science book series (LNCS, volume 667)


This paper addresses the problem of classifying observations when features are context-sensitive, specifically when the testing set involves a context that is different from the training set. The paper begins with a precise definition of the problem, then general strategies are presented for enhancing the performance of classification algorithms on this type of problem. These strategies are tested on two domains. The first domain is the diagnosis of gas turbine engines. The problem is to diagnose a faulty engine in one context, such as warm weather, when the fault has previously been seen only in another context, such as cold weather. The second domain is speech recognition. The problem is to recognize words spoken by a new speaker, not represented in the training set. For both domains, exploiting context results in substantially more accurate classification.


  1. 1.
    D.W. Aha, D. Kibler, and M.K. Albert: Instance-based learning algorithms, Machine Learning, 6, 37–66,1991.Google Scholar
  2. 2.
    D. Kibler, D.W. Aha, and M.K. Albert: Instance-based prediction of real-valued attributes, Computational Intelligence, 5, 51–57, 1989.Google Scholar
  3. 3.
    B.V. Dasarathy: Nearest Neighbor Pattern Classification Techniques, (edited collection), Los Alamitos, CA: IEEE Press, 1991.Google Scholar
  4. 4.
    N.R. Draper and H. Smith: Applied Regression Analysis, (second edition), New York, NY: John Wiley & Sons, 1981.Google Scholar
  5. 5.
    S.E. Fahlman and C. Lebiere: The Cascade-Correlation Learning Architecture, (technical report), CMU-CS-90-100, Pittsburgh, PA: Carnegie-Mellon University, 1991.Google Scholar
  6. 6.
    A.J. Katz, M.T. Gately, and D.R. Collins: Robust classifiers without robust features, Neural Computation, 2, 472–479,1990.Google Scholar
  7. 7.
    P.D. Turney and M. Halasz: Contextual normalization applied to aircraft gas turbine engine diagnosis, (in press), Journal of Applied Intelligence, 1993.Google Scholar
  8. 8.
    D. Deterding: Speaker Normalization for Automatic Speech Recognition, (Ph.D. thesis), Cambridge, UK: University of Cambridge, Department of Engineering, 1989.Google Scholar
  9. 9.
    A.J. Robinson: Dynamic Error Propagation Networks, (Ph.D. thesis), Cambridge, UK: University of Cambridge, Department of Engineering, 1989.Google Scholar
  10. 10.
    P.M. Murphy and D.W. Aha: UCI Repository of Machine Learning Databases, Irvine, CA: University of California, Department of Information and Computer Science, 1991.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1993

Authors and Affiliations

  • Peter D. Turney
    • 1
  1. 1.Knowledge Systems Laboratory, Institute for Information TechnologyNational Research Council CanadaOttawaCanada

Personalised recommendations