Abstract
We propose a fast feature selection method in supervised learning for multi-valued attributes. The main idea is to rewrite the multi-valued problem in the space of examples into a boolean problem in the space of pairwise examples. On basis of this approach, we can use point correlation coefficient which is null in the case of conditional independence, and verifies a formula connecting partial coefficients with marginal coeffcients. This property allows to reduce considerably the computing times because a single pass over the database is necessary to compute all coeffcients. We test our algorithm on benchmark databases.
Chapter PDF
Similar content being viewed by others
References
D. Aha. Tolerating noisy, irrelelvant and novel attributes in instance-based algorithms. International Journal of Man-Machine Studies, 36:267–287, 1992.
R. Battiti. Using mutual information for selecting features in supervised neural net learning. IEEE Transactions on Neural Networks, 5(4):537–550, 1994.
S.D. Bay. The uci kdd archive [http://kdd.ics.uci.edu]. Irvine,CA: University of California, Department of Computer Science, 1999.
A. Blum and P. Langley. Selection of relevant feature and examples in machine learning. Artificial Intelligence, pages 245–271, 1997.
J.J. Daudin. Analyse factorielle des dependances partielles. Revue de Statistique Appliquee, 29(2):15–29, 1981.
J. Dougherty, R. Kohavi, and M. Sahami. Supervised and unsupervised discretization of continuous attributes. In Morgan Kaufmann, editor, Machine Learning: Proceedings of the 12th International Conference (ICML-95), pages 194–202, 1995.
U. Fayyad, G. Piatetsky-Shapiro, and P. Smyth. Knowledge discovey and data mining: Towards an unifying framework. In Proceedings of the 2nd International Conference on Knowledge Discovery and Data Mining, 1996.
Eibe Frank and Ian H. Witten. Making better use of global discretization. In Proc.16th International Conf. on Machine Learning, pages 115–123. Morgan Kaufmann, San Francisco, CA, 1999.
L.A. Goodman and W.H. Kruskall. Measures of association for cross classifcations.Journal of American Statistical Association, 49:732–764, 1954.
R.C. Holte. Very simple classification rules perform well on most commonly used datasets. Machine Learning, 11:63–91, 1993.
G. John and P. Langley. Static versus dynamic sampling for data mining. In Proceedings of the 2nd International Conference on Knowledge Discovery in Databases and Data Mining. AAAI/MIT Press, 1996.
R. Kohavi and G. John. Wrappers for feature subset selection. Journal of Artificial Intelligence, Special issue on Relevance, 1997.
S. Lallich and R. Rakotomalala. Les entropies de rangs généralisés en induction par arbres. In Proceedings of 7 mes Journées de la Société Francophone de Classification-SFC’99, pages 101–107, September 1999.
K-C. Lee. A technique of dynamic feature selection using the feature group mutual information. In Proceedings of the Third PAKDD-99, pages 138–142, 1999.
I.C. Lerman. Correlation partielle dans le cas qualitatif. Technical Report 111, INRIA, 1982.
H. Liu and H. Motoda. Feature Extraction, Construction and Selection: A Data Mining Perspective. Kluwer Academic Publishers, 1998.
H. Liu and H. Motoda. Feature selection for knowledge discovery and data mining, volume 454 of The kluwer international series in engineering and computer science. Kluwer, 1998.
F. Marcotorchino. Utilisation des comparaisons par paires en statistique des contingences-partie iii. Technical Report F 081, Centre Scientifique IBM-France, 1985.
D. Michaud. Filtrage et Selection D’attributs En Apprentissage. PhD thesis, Universite de Franche-Comte, 1999.
M. Olszak and G. Ritschard. The behaviour of nominal and ordinal partial association measures. The statistician, 44(2):195–212, 1995.
R. Rakotomalala and S. Lallich. Sélection rapide de variables booléennes en apprentissage supervisé. In Proceedings of 2nd Conférence Apprentissage-CAP’2000, pages 225–234, 2000.
R. Rakotomalala, S. Lallich, and S. Di Palma. Studying the behavior of generalized entropy in induction trees using a m-of-n concept. In Proceedings of the Third European Conference PKDD’99, pages 510–517, 1999.
G. Saporta. Liaisons entre plusieurs ensembles de variables et codage de données qualitatives. PhD thesis, 1975.
G. Saporta. Quelques applications des operateurs d’Escouffier au traitement des variables qualitatives. Statistique et Analyse de Donnees, (1):38–46, 1976.
D.A. Zighed, S. Rabaseda, and R. Rakotomalala. Fusinter: a method for discretization of continuous attributes for supervised learning. International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 6(33):307–326, 1998.
-D.A. Zighed and R. Rakotomalala. Graphes d’Induction-Apprentissage et Data Mining. Hermes, 2000.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2000 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Lallich, S., Rakotomalala, R. (2000). Fast Feature Selection using Partial Correlation for Multi-valued Attributes. In: Zighed, D.A., Komorowski, J., Żytkow, J. (eds) Principles of Data Mining and Knowledge Discovery. PKDD 2000. Lecture Notes in Computer Science(), vol 1910. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-45372-5_22
Download citation
DOI: https://doi.org/10.1007/3-540-45372-5_22
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-41066-9
Online ISBN: 978-3-540-45372-7
eBook Packages: Springer Book Archive