论文部分内容阅读
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.
Since the concept of rough set was proposed by Pawlak, people are always interested in the approximate accuracy of rough set, and there are many literatures about the accuracy of approximation.In rough set theory, the precision is to quantify the inaccuracy caused by rough set boundary Based on the analysis of the approximate accuracy of the traditional accuracy and the excess entropy based on the equivalence graph, a new definition of precision is proposed, and the accuracy of the new definition is found to be more reasonable by comparison. At the same time, The new definition of precision is applied to the attribute reduction. By comparing the examples, it is found that the attribute reduction proposed in this paper is more feasible.