A hybrid feature selection approach based on the bayesian network classifier and rough sets

Abstract

The paper proposes a hybrid feature selection approach based on Rough sets and Bayesian network classifiers. In the approach, the classification result of a Bayesian network is used as the criterion for the optimal feature subset selection. The Bayesian network classifier used in the paper is a kind of naive Bayesian classifier. It is employed to implement classification by learning the samples consisting of a set of texture features. In order to simplify feature reduction using Rough Sets, a discrete method based on C-means clustering method is also presented. The proposed approach is applied to extract residential areas from panchromatic SPOT5 images. Experiment results show that the proposed method not only improves classification quality but also reduces computational cost. © 2008 Springer-Verlag Berlin Heidelberg.