H. Grosskreutz, and S. Rüping. Data Mining and Knowledge Discovery, 19 (2):
210--226(October 2009)
Abstract
Abstract Subgroup discovery is a Knowledge Discovery task that aims at finding subgroups of a population with high generality and distributional
unusualness. While several subgroup discovery algorithms have been presented in the past, they focus on databases with nominalattributes or make use of discretization to get rid of the numerical attributes. In this paper, we illustrate why the replacementof numerical attributes by nominal attributes can result in suboptimal results. Thereafter, we present a new subgroup discoveryalgorithm that prunes large parts of the search space by exploiting bounds between related numerical subgroup descriptions.The same algorithm can also be applied to ordinal attributes. In an experimental section, we show that the use of our newpruning scheme results in a huge performance gain when more that just a few split-points are considered for the numericalattributes.
%0 Journal Article
%1 henrik2009subgroup
%A Grosskreutz, Henrik
%A Rüping, Stefan
%D 2009
%J Data Mining and Knowledge Discovery
%K continuous ecml09 imported sg_Discovery
%N 2
%P 210--226
%T On subgroup discovery in numerical domains
%U http://dx.doi.org/10.1007/s10618-009-0136-3
%V 19
%X Abstract Subgroup discovery is a Knowledge Discovery task that aims at finding subgroups of a population with high generality and distributional
unusualness. While several subgroup discovery algorithms have been presented in the past, they focus on databases with nominalattributes or make use of discretization to get rid of the numerical attributes. In this paper, we illustrate why the replacementof numerical attributes by nominal attributes can result in suboptimal results. Thereafter, we present a new subgroup discoveryalgorithm that prunes large parts of the search space by exploiting bounds between related numerical subgroup descriptions.The same algorithm can also be applied to ordinal attributes. In an experimental section, we show that the use of our newpruning scheme results in a huge performance gain when more that just a few split-points are considered for the numericalattributes.
@article{henrik2009subgroup,
abstract = {Abstract Subgroup discovery is a Knowledge Discovery task that aims at finding subgroups of a population with high generality and distributional
unusualness. While several subgroup discovery algorithms have been presented in the past, they focus on databases with nominalattributes or make use of discretization to get rid of the numerical attributes. In this paper, we illustrate why the replacementof numerical attributes by nominal attributes can result in suboptimal results. Thereafter, we present a new subgroup discoveryalgorithm that prunes large parts of the search space by exploiting bounds between related numerical subgroup descriptions.The same algorithm can also be applied to ordinal attributes. In an experimental section, we show that the use of our newpruning scheme results in a huge performance gain when more that just a few split-points are considered for the numericalattributes.},
added-at = {2009-09-24T11:02:21.000+0200},
author = {Grosskreutz, Henrik and Rüping, Stefan},
biburl = {https://www.bibsonomy.org/bibtex/2072a634dda935bcd13132e8944a8ba3c/lemmi},
description = {SpringerLink - Zeitschriftenbeitrag},
interhash = {0c722431ed8bbb2c3482f99750699f10},
intrahash = {072a634dda935bcd13132e8944a8ba3c},
journal = {Data Mining and Knowledge Discovery},
keywords = {continuous ecml09 imported sg_Discovery},
month = {#oct#},
number = 2,
pages = {210--226},
timestamp = {2009-09-24T11:02:21.000+0200},
title = {On subgroup discovery in numerical domains},
url = {http://dx.doi.org/10.1007/s10618-009-0136-3},
volume = 19,
year = 2009
}