启发式值约简算法的研究与实现

The Study and Implementation of Heuristic Value Reduction

摘要: 在粗糙集理论的基础上，本文研究了启发式值约简的过程。本文研究的就是在属性约简完成后的启发式值约简算法，它一般先构造由约简属性集合和决策属性组成的决策表；然后利用一定的启发式信息对其进行去重操作，得到的新表将作为值约简的初始决策表；最后，判断记录中的各个属性值是否冗余，删除冗余属性值，对记录的属性值进行约简得到近似最小规则集。最终实现了其测试系统。

Abstract: Based on the research of rough set theory, this paper studies the process of heuristic value reduction. It usually constructs the decision table composed of the reduced attribute set and the decision attribute. Then, the heuristic information is used to perform the de-operation and delete the duplicate information. Finally, the new decision table works as the initial decision table and heuristic algorithm is used to judge whether the attribute values in the records are redundant or necessary. The redundant attribute values are deleted and the attribute values of the records are reduced to get the approximate minimum rule set. At last the test system is implemented.