基于精度的属性约简新方法

自Pawlak提出粗糙集概念以来,人们就一直对粗糙集的近似精度很感兴趣,出现了不少有关近似精度的文献.在粗糙集理论中,精度是量化由粗糙集边界引起的不精确性的一种重要数字特征.在分析传统精度和基于等价关系图的过剩熵的近似精度的基础上,提出了一种新的精度定义.比较发现,新定义的精度更具有合理性.同时把这个新定义的精度运用到了属性约简上,通过实例比较发现,本文提出的属性约简更具有可行性.

A New Method of Attribute Reduction Based on the Precision

Many people have been interested in the concept of accurcy of approximation since the rough sets were introduced by Pawlak.Many papers about the accuracy of approximation are appeared.In rough set theory,accuracy measure is an important numerical characterization that quantifies the imprecision of a rough set caused by its boundary region.By comparing with the Pawlak accuracy of approximation and the based on excessentropy accuracy of approximation,a new accuracy of approximation for the rough sets is introduced in this paper.The advantage of the new definition can be seen.At the same time apply the new definition on the attribute reduction.Through examples,we can see that the new attribute reduction way is more feasible.