| 张 杰1，张燕兰1*，林艺东2.基于图的覆盖决策信息系统属性约简新算法[J].海南师范大学学报自科版,2022,35(1):16-24 |
| 基于图的覆盖决策信息系统属性约简新算法 |
| A New Attribute Reduction Algorithm for Covering DecisionInformation System Based on Graph |
| |
| DOI：10.12051/j.issn.1674-4942.2022.01.003 |
| 中文关键词: 属性约简 顶点覆盖 图论 覆盖粗糙集 关联矩阵 |
| 英文关键词: attribute reduction vertex cover graph theory covering roughset incidence ｍatrix |
| 基金项目:福建省自然科学基金项目（2019J01749，2019J01748） |
|
| 摘要点击次数: 547 |
| 全文下载次数: 812 |
| 中文摘要: |
| 属性约简是粗糙集理论的重要研究内容之一，主要是用于去除高维数据的冗余信息。
利用图论求约简是覆盖决策信息系统属性约简的一个新方法，其将覆盖决策系统的约简等价于超
图的极小顶点覆盖。本研究提出一种新的覆盖决策信息系统的属性约简算法，该算法采用三步策
略：首先确定覆盖决策信息系统的辨识集，然后得到超图的关联矩阵，最后用贪心法求出超图的极
小顶点覆盖。实验结果表明，新的属性约简算法能够有效地降低数据维数和时间复杂度。 |
| 英文摘要: |
| :At present, attribute reduction is one of the important research issues for rough set theory and is mainly used to
remove redundant information of high-dimensional data. It is proved that solving the attribute reduction of covering deci⁃
sion information system is equivalent to finding the minimal vertex covering of a hypergraph. Our new attribute reduction al⁃
gorithm was proposed with a three-step strategy. Firstly, the discernibility set of the covering decision information system
was determined, then the incidence matrix of a hypergraph was specified and finally the minimal vertex cover of the hyper⁃
graph with greedy method was found. Experimental results showed that the new attribute reduction algorithm can reduce the
dimension of data as well as the time complexity effectively. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
