Abstract
Lately, decision trees became an important tool in marketing research, e.g. to classify potential customers in order to send individual ads to them. In contrast to many other situations here the classification should be done in such a manner that not the inhomogeneity of subsets is minimised but a cost orientated loss function. Unfortunately, split criteria based on cost orientated loss functions have the disadvantage to be too often indifferent between different splits. This is among others due to the local optimality of the split calculations ([1]) and due to the facts that a further improvement of the loss function could be better achieved if the best split in a “class of indifferent or nearly indifferent splits” is chosen. This paper presents a parametric enlargement of the local search for optimal splits using cost orientated loss functions.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Breiman, L.; Friedman, J.H.; Olshen, R.A.; Stone, C.J. (1984) Classification and Regression Trees. Wadsworth Inc., Pacific Grove, California.
Hippner, H.; Küsters, U.; Meyer, M.; Wilde, K. (2001) Handbuch Datamining im Marketing. Vieweg, Gabler, Braunschweig, Wiesbaden.
Kass, G.V. (1980) An Exploratory Technique for Investigating Large Quantities of Categorial Data. In: Applied Statistics. Vol. 29, No. 2, S. 119–127
Quinlan, J.R. (1986) Introduction of decision trees. In: Machine Learning 1(1), S. 81–106
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2002 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Hilbert, A., Paul, H. (2002). Decision Trees with Parametric Enlarged Local Search. In: Chamoni, P., Leisten, R., Martin, A., Minnemann, J., Stadtler, H. (eds) Operations Research Proceedings 2001. Operations Research Proceedings 2001, vol 2001. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-50282-8_11
Download citation
DOI: https://doi.org/10.1007/978-3-642-50282-8_11
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-43344-6
Online ISBN: 978-3-642-50282-8
eBook Packages: Springer Book Archive