重庆邮电大学学报（自然科学版）2012，Vol.24Issue(4)：490-494,5.DOI:10.3979/j.issn.1673-825X.2012.04.019
一种基于二进制分辨矩阵的属性约简新算法
New algorithm for attribute reduction based on binary discernibility matrix
摘要
Abstract
The attribute reduction of decision tables is one of the key issues in rough set theory. Effective and efficient heuristic reduction algorithms are deserved to be further studied. Herein, a concept of " weighted importance" is defined based on binary discernibility matrix to heuristically measure the significances of attributes of a decision table. The concept comprehensively makes good use of the characteristics of columns and rows of binary discernibility matrix and integrates them into a normalized concept; consequently, independent sorting computation in respective directions of column and row is effectively avoided. Then a new heuristic algorithm for attribute reduction is proposed based on concept of "weighted importance" . Simulation experiment results show that compared with some benchmark algorithms, the new algorithm runs faster and is more possible to get optimal reductions of decision tables.关键词
粗糙集理论/决策表/属性约简/二进制分辨矩阵/加权重要度Key words
rough set theory /decision table/attribute reduction/binary discernibility matrix/weighted importance分类
信息技术与安全科学引用本文复制引用
赵军,陈宸..一种基于二进制分辨矩阵的属性约简新算法[J].重庆邮电大学学报（自然科学版）,2012,24(4):490-494,5.基金项目
重庆市自然科学基金(2008BA2017) （2008BA2017）
重庆市教委科学技术研究基金(KJ080510,KJ110522) （KJ080510,KJ110522）
重庆市杰出青年科学基金(2008BA2041) （2008BA2041）
