文章基本信息

标题：Solving Attribute Reduction Problem using Wrapper Genetic Programming
本地全文：下载
作者：Mohammed Alweshah ; Omar A. Alzubi ; Jafar A. Alzubi 等
期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2016
卷号：16
期号：5
页码：78-84
出版社：International Journal of Computer Science and Network Security
摘要：Attribute reduction (AR) represents a NP-hard problem, and it is be identified as the problematic issue of pinpointing the least (possible) subset of characteristics taken from the reference set. The key issue related to characteristics selectors is the production of a minimal number of reductions representing the reliable meaning of all characteristics. Nevertheless, there is no approach that can ensure optimality in the process of solving this issue. However, some methods are more efficient compared to other ones due to some characteristics of the algorithm. The research of this thesis aims at providing efficient ways that help us find the characteristics which are known as the most informative ones and the least possible features with least possible data loss. This has been done through the combination between wrapper approach and genetic programming algorithm, Wrapper Genetic Programming (WGP). Numerical experiment carried out on 10 real word dataset from the University of California Irvine benchmark data sets (UCI) Repository of Machine Learning Databases has been used and presented in order to show that WGP can give competitive solutions in an efficient manner compared to approaches available in the literature on this issue.
关键词：Classification; attribute reduction; genetic programing algorithm; wrapper approach.