计算机技术与发展2018，Vol.28Issue(4)：99-103,5.DOI:10.3969/j.issn.1673-629X.2018.04.021
基于相对辨识关系的属性约简算法
An Attribute Reduction Algorithm Based on Relative Discernible Relation
摘要
Abstract
We mainly introduce some related concepts in decision information system,such as discernible relation and relative discernible relation,and complete the attribute set independence judgment.Connecting the discernibility and relative discernibility of the attribute set with the object of identification of the property set,we also study the evaluation methods of attribute importance between the two kinds of relative identification and finish the judgment of attribute's independence or dependence and whether to do reduction of decision informa-tion system.Based on the relative recognition relation,the corresponding improvement algorithm is given,by which the variation of the relative recognition caused by the addition or subtraction of attribute set attributes is described.The algorithm determines the relatively dis-cernible relation of each attribute from the condition attribute set,taking the largest number of object as the reduction set where the attrib-utes are gradually added until meeting the conditions of reduction.It is a kind of seedless attribute reduction algorithm,which is reduced to a certain extent regardless of time complexity or reduction workload.Its validity is verified by an example.关键词
粗糙集/相对可辨识关系/属性集/属性约简Key words
rough sets/relative discernible relation/attribute set/attribute reduction分类
信息技术与安全科学引用本文复制引用
孙滨,毛建景..基于相对辨识关系的属性约简算法[J].计算机技术与发展,2018,28(4):99-103,5.基金项目
河南省重大科技专项(162102210319) （162102210319）
河南省基础与前沿技术研究计划项目(142300410283) （142300410283）
河南省科学技术研究重点项目(12B520063,14B520065) （12B520063,14B520065）
河南省高等学校青年骨干教师资助计划项目(2016GGJS-182) （2016GGJS-182）
