一种基于Rough集理论的属性约简启发式算法

属性约简是知识发现中的关键问题之一．为了能够有效地获取决策表中属性的最小相对约简，在Rough集理论的基础上构造了一个新的算子，将信息论角度定义的属性的重要性作为启发式信息，来描述在决策表中条件属性所提供的知识对决策属性的影响；并采用宽度优先搜索策略，提出了一种新的属性约简启发式算法．以原始条件属性集为起点并结合算子，通过向属性核的递减式逼近，得到属性的最小相对约简．实例分析表明，该算法能有效地对决策表属性进行约简．

Heuristic Algorithm for Reduction of Attributes Based on Rough Set Theory

Attribute reduction is one of the key problems in the knowledge discovery. Based on the rough set theory, a new operator is constructed to effectively achieve the minimal relative reduction of attribute in the decision table. Regarding the significance of attributes defined from the viewpoint of information theory as heuristic information and applying the breadth-first search strategy, a new heuristic algorithm for reduction of attribute is proposed. Acquiring optimal relative reduction by descending approach to core of attribute from original set of conditional attribute and combining with operator. Finally, the experimental results show that this algorithm was effective in attributes reduction of decision tables.