一种基于Rough集的属性值约简算法  

A Rough Set-based Algorithm for Attribute Value Reduction

作　　者：胡斐[1] 张峰筠[1] 刘少辉[2] 

机构地区：[1]上海体育学院体育管理系,上海200438 [2]中国科学院计算技术研究所智能信息处理重点实验室,北京100080

出　　处：《计算机工程与应用》2003年第31期48-51,共4页Computer Engineering and Applications

基　　金：国家自然科学基金资助项目(编号:60073019)

摘　　要：文章将Rough集理论应用于不同类型的决策表(一致决策表和不一致决策表)的约简,给出了广义决策、决策规则的一致程度、属性值重要性等定义,在此基础上提出了一种基于Rough集的属性值约简算法。该算法不仅能得到更为简洁的决策规则,而且能保持决策规则的一致程度不变。实例分析表明该算法是可行的。Rough set theory is applied to the reduction of different types of decision table(consistent decision table and inconsistent decision table).A lot of definitions such as the generalized decision,the degree of a decision rule's consistency and the significance of attribute value are given.Based on these definitions,a rough set-based algorithm for attribute value reduction is proposed.Not only can it get more concise decision rules,but also it can keep the degree of consistency unchangeable.The analysis of an example shows that the algorithm is feasible.

关 键 词：ROUGH集 决策表 值核 值约简 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]