A preliminary study on the use of fuzzy rough set based feature selection for improving evolutionary instance selection algorithms

Authors:
Joaquín Derrac;Chris Cornelis;Salvador García;Francisco Herrera
Affiliations:
Dept. of Computer Science and Artificial Intelligence, CITIC-UGR, University of Granada, Granada, Spain;Dept. of Applied Mathematics and Computer Science. Ghent University, Gent, Belgium;Dept. of Computer Science. University of Jaén, Jaén, Spain;Dept. of Computer Science and Artificial Intelligence, CITIC-UGR, University of Granada, Granada, Spain
Venue:
IWANN'11 Proceedings of the 11th international conference on Artificial neural networks conference on Advances in computational intelligence - Volume Part I
Year:
2011

Citing 8
Cited 0

Instance Selection and Construction for Data Mining

Instance Selection and Construction for Data Mining
Introduction to Evolutionary Computing

Introduction to Evolutionary Computing
Computational Methods of Feature Selection (Chapman & Hall/Crc Data Mining and Knowledge Discovery Series)

Computational Methods of Feature Selection (Chapman & Hall/Crc Data Mining and Knowledge Discovery Series)
A memetic algorithm for evolutionary prototype selection: A scaling up approach

Pattern Recognition
Improved heterogeneous distance functions

Journal of Artificial Intelligence Research
Attribute selection with fuzzy decision reducts

Information Sciences: an International Journal
IFS-CoCo: Instance and feature selection based on cooperative coevolution with nearest neighbor rule

Pattern Recognition
Using evolutionary algorithms as instance selection for data reduction in KDD: an experimental study

IEEE Transactions on Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

In recent years, the increasing interest in fuzzy rough set theory has allowed the definition of novel accurate methods for feature selection. Although their stand-alone application can lead to the construction of high quality classifiers, they can be improved even more if other preprocessing techniques, such as instance selection, are considered. With the aim of enhancing the nearest neighbor classifier, we present a hybrid algorithm for instance and feature selection, where evolutionary search in the instances' space is combined with a fuzzy rough set based feature selection procedure. The preliminary results, contrasted through nonparametric statistical tests, suggest that our proposal can improve greatly the performance of the preprocessing techniques in isolation.