变精度集对粗糙集模型中的属性约简

属性约简是粗糙集理论的核心内容之一。论文是继续文献[8]的工作,在变精度集对粗糙集模型的基础上,定义了变精度的重要性算子和变精度的近似约简等概念,并由此给出了一种属性约简的启发式算法。算法既能保证属性约简的准确性,又能增加其灵活性,它可以通过对相似度α和精度β的调节,按照广度优先搜索策略,从条件属性集中逐一删除重要性最小的属性,从而得到一个满足相似度和精度要求的近似约简。同时,它也是完备信息系统的属性约简算法的推广(当α=1,β=0时)。最后通过一个实例,分析说明算法的可行性和有效性。

An Algorithm for Attributes Reduction in Variable Precision Rough Set Model Based on Set Pair Analysis

Attribute reduction is one of the key problems in rough set theory.This paper continues the research into the variable precision rough set model based on set pair analysis.By introducing some concepts such as significance operator of attributes and approximate reduction,a new algorithm for attribute reduction is proposed.Not only has the accuracy of attribute reduction been guaranteed,but also its flexibility has also been increased by adjusting the values of the similarity α and the relatively classified fault rate β.In addition,the new heuristic algorithm for attribute reduction in this paper is the generalization of the algorithm in complete information system.Finally an example shows that this new algorithm is both feasible and effective in practice.