On fuzzy-rough sets approach to feature selection

In this paper, we have shown that the fuzzy-rough set attribute reduction algorithm [Jenson, R., Shen, Q., 2002. Fuzzy-rough sets for descriptive dimensionality reduction. In: Proceedings of IEEE International Conference on Fuzzy Systems, FUZZ-IEEE’02, May 12–17, pp. 29–34] is not convergent on many...

Full description

Saved in:
Bibliographic Details
Published inPattern recognition letters Vol. 26; no. 7; pp. 965 - 975
Main Authors Bhatt, Rajen B., Gopal, M.
Format Journal Article
LanguageEnglish
Published Elsevier B.V 15.05.2005
Subjects
Computational complexity
Feature selection
Fuzzy sets
Rough sets
Fuzzy sets
Feature selection
Computational complexity
Rough sets
Online AccessGet full text

Cover

More Information
Summary:In this paper, we have shown that the fuzzy-rough set attribute reduction algorithm [Jenson, R., Shen, Q., 2002. Fuzzy-rough sets for descriptive dimensionality reduction. In: Proceedings of IEEE International Conference on Fuzzy Systems, FUZZ-IEEE’02, May 12–17, pp. 29–34] is not convergent on many real datasets due to its poorly designed termination criteria; and the computational complexity of the algorithm increases exponentially with increase in the number of input variables and in multiplication with the size of data patterns. Based on natural properties of fuzzy t-norm and t-conorm, we have put forward the concept of fuzzy-rough sets on compact computational domain, which is then utilized to improve the computational efficiency of FRSAR algorithm. Speed up factor as high as 622 have been achieved with this concept with improved accuracy. We also restructure the algorithm with efficient termination criteria to achieve the convergence on all the datasets and to improve the reliability of selected set of features.
ISSN:0167-8655
1872-7344
DOI:10.1016/j.patrec.2004.09.044