Feature Selection on the Basis of Rough Set Theory and Univariate Marginal Distribution Algorithm

DOI

10.2991/ammsa-17.2017.83How to use a DOI?

Keywords

feature selection; rough set theory; UMDA

Abstract

Feature selection is an important preprocessing step in machine learning. The aim of feature selection is to find an optimal subset from original features that satisfies a criterion. Rough set theory (RST) is one of the most effective ways to solve feature selection problem, but RST is inefficient in large scale datasets. In order to solve this problem, in this paper, we proposed a novel feature selection algorithm RSUMDA on the basis of univariate marginal distribution algorithm. RST was used to obtain the significance of each feature as the original probability of UMDA and then UMDA was to search the optimal feature subset that using the number of the selected feature and the accuracy of the classifier as fitness function. Experimentation was carried out in 4 UCI datasets. The results showed that our algorithm could effectively reduce the number of the features, improve the accuracy of the classifier and quicken the convergence rate.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Bin Wei
AU  - Minqing Zhang
AU  - Longfei Liu
AU  - Jing Zhao
PY  - 2017/05
DA  - 2017/05
TI  - Feature Selection on the Basis of Rough Set Theory and Univariate Marginal Distribution Algorithm
BT  - Proceedings of the 2017 International Conference on Applied Mathematics, Modelling and Statistics Application (AMMSA 2017)
PB  - Atlantis Press
SP  - 369
EP  - 372
SN  - 1951-6851
UR  - https://doi.org/10.2991/ammsa-17.2017.83
DO  - 10.2991/ammsa-17.2017.83
ID  - Wei2017/05
ER  -