Abstract
Supervised classification uses a training set to construct a classifier such as a decision tree. Normally, the training set is discarded once the training process is complete. By imprinting information about the training population onto the classifier, we can make use of the extrema at each node as “canaries”, warning us that we have left the well explored area of parameter space and have crossed into a domain where the classifier is unreliable. This technique can identify training set deficiencies; provide reliability estimates for decision tree classifiers; improve the results of multi-tree voting; and provide helpful visualization tools. See http://www-gsss.stsci.edu/PublishedPapers/Canaries_SCMA.htm for the poster version of this paper.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Heath, D., et al., 1996, Cognitive Technology: In Search of a Human Interface, eds. Gorayska & Mey.
Laidler, V. G., et al., 1996, Bull.Am.Astron.Soc., 188, 5421
Murthy, S. K., Kasif, S., & Salzberg, S. 1994, J. Artificial Intelligence Research, 2, 1
White, R. L. 1997, Statistical Challenges in Modern Astronomy II, eds. G. J. Babu & E. D. Feigelson (New York: Springer), p. 135.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2003 Springer-Verlag New York, Inc.
About this paper
Cite this paper
Laidler, V.G., White, R.L. (2003). Canaries in the Data Mine: Improving Trained Classifiers. In: Statistical Challenges in Astronomy. Springer, New York, NY. https://doi.org/10.1007/0-387-21529-8_49
Download citation
DOI: https://doi.org/10.1007/0-387-21529-8_49
Publisher Name: Springer, New York, NY
Print ISBN: 978-0-387-95546-9
Online ISBN: 978-0-387-21529-7
eBook Packages: Springer Book Archive