A neuro fuzzy algorithm for feature subset selection

Citation
B. Chakraborty et G. Chakraborty, A neuro fuzzy algorithm for feature subset selection, IEICE T FUN, E84A(9), 2001, pp. 2182-2188
Citations number
13
Language
INGLESE
art.tipo
Article
Categorie Soggetti
Eletrical & Eletronics Engineeing
Journal title
IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES
ISSN journal
0916-8508 → ACNP
Volume
E84A
Issue
9
Year of publication
2001
Pages
2182 - 2188
Database
ISI
SICI code
0916-8508(200109)E84A:9<2182:ANFAFF>2.0.ZU;2-#
Abstract
Feature subset selection basically depends on the design of a criterion fun ction to measure the effectiveness of a particular feature or a feature sub set and the selection of a search strategy to find out the best feature sub set. Lots of techniques have been developed so far which are mainly categor ized into classifier independent falter approaches and classifier dependant wrapper approaches. Wrapper approaches produce good results but are comput ationally unattractive specially when nonlinear neural classifiers with com plex learning algorithms are used. The present work proposes a hybrid two s tep approach for finding out the best feature subset from a large feature s et in which a fuzzy set theoretic measure for assessing the goodness of a f eature is used in conjunction with a multilayer perceptron (MLP) or fractal neural network (FNN) classifier to take advantage of both the approaches. Though the process does not guarantee absolute optimality, the selected fea ture subset produces near optimal results for practical purposes. The proce ss is less tune consuming and computationally light compared to any neural network classifier based sequential feature subset selection technique. The proposed algorithm has been simulated with two different data sets to just ify its effectiveness.