一种基于知识粒度单调性的属性约简算法  

AN ATTRIBUTE REDUCTION ALGORITHM BASED ON KNOWLEDGE GRANULARITY MONOTONICITY

作　　者：王晓宇[1] 徐章艳[1] 张伟[1] 

机构地区：[1]广西师范大学计算机科学与信息工程学院,广西桂林541004

出　　处：《计算机应用与软件》2013年第11期38-41,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(60963008);广西自然科学基金项目(2011GXNSFA018163);广西研究生教育创新计划项目

摘　　要：信息系统的属性约简和求核算法是粗糙集目前研究的主要内容。除正区域、差别矩阵、区分对象对之外,知识粒度也适于进行属性约简和求核计算。根据知识粒度的单调性,以知识粒度代数式为启发信息设计属性约简算法,达到降低时间复杂度的目的。最后用实例说明该算法的有效性,并得出该算法的时间复杂度为O(|A|2|U|)。Attribute reduction and core calculation algorithm in information system are the main contents to be studied in rough set at present. In addition to positive region, discernibility matrix and discernibility object pairs, the knowledge granularity is also suitable for attribute reduction and core calculation. In this paper, according to the monotonicity of knowledge granularity, we design the attribute reduction algorithm for heuristic information using the algebraic expression of knowledge granularity to achieve the goal of reducing the time complexity. An example is given to illustrate the validity of the algorithm at the end, the time complexity of the algorithm 0（｜A｜^2｜U｜） is presented as well.

关 键 词：粗糙集 信息系统 知识粒度 属性约简 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]