基于属性重要性的启发式属性约简算法

针对粗糙集理论中的属性约简问题做了探讨研究。从寻找属性约简的角度,首先描述了决策表中的属性的重要性,并利用已求得的正区域使处理数据的范围不断缩小,约简集中的属性从核集开始,通过向属性核添加重要性最大的属性,得到属性的最小相对约简。从而减少求约简的时间。最后进行实证,该算法同传统的算法相比,在计算量减少的同时能得到更简约的结果,证明了该算法的正确性和可行性。

Heuristic Reduction Algorithm Based on Attribute Importance

This article has done the research on attribute reduction of rough set theory.This paper is from the perspective of looking for attribute reduction,This paper first described the properties importance in the decision-making table and use the positive region which has been obtained,narrowed the scope of data processing,attribute reduction begin with core set,through adding the attributes which are the most important to properties nuclear,get the smallest relative reduction of properties.Thereby it reduced the time of demanding reduction.Finally,an example of a complete demonstration of the method,confirmed that the algorithm with the traditional method,in calculating the amount of the reduction at the same time can be more simple result The results proved that the correctness and feasibility of the algorithmis.