变精度粗糙集论文.doc

摘要 粗糙集理论是上世纪八十年代初由波兰数学家Pawlak提出的一种刻划不完整性和不确定性的数学工具能有效地分析不精确不一致(inconsistent)不完整(incomplete)等各种不完备的信息Pawlak粗糙集理论的一种扩充，它通过设置阈值参数（），放松了对标准粗糙集理论近似边界的严格定义，增强了粗糙集模型的抗干扰能力。 本文对基于变精度粗糙集的属性约简算法进行了改进。总结了现有方法，并对现有文献提出的基于可分辨矩阵的粗糙集属性约简算法进行改进，进而提出了一个新的基于变精度粗糙集的可分辨矩阵的属性约简算法——VMAR算法。文章介绍了粗糙集和变精度粗糙集的基本理论，分析了变精度粗糙集模型中β值的取值对约简效果的影响，最后通过UCI数据库进行实验，验证了改进的VMAR算法的优势。 关键词：粗糙集；精度粗糙集；属性约简；阈值 Abstract rough set theory is in the early 1980s by the Polish mathematician Pawlaks not a characterization of the integrity and uncertainties mathematical tools to effectively analysis inaccurate, inconsistent (incon sisten t), incomplete (Incomp lete), and other incomplete information. The main idea is to use the knowledge base of known, imprecise or uncertainty of knowledge known by the knowledge base of knowledge in the (approximate) characterization. Variable precision rough set Pawlak is a rough set of expansion, It passed the threshold set parameters (), to relax the standard rough set theory similar to the strict definition of the border, strengthen the rough set model of anti-jamming ability. This article is based on variable precision rough set of attribute reduction algorithm improvements. Summed up the existing methods, and the literature [16] can be resolved based on the matrix of rough set attribute reduction algorithm to improve, then proposed a new variable precision rough set based on the matrix can distinguish between the attribute reduction algorithm -- VMAR algorithm.The article first introduced rough sets and variable precision rough set some of the basic theory, and then prove that the variable precision rough set model of β value of the value of the reduction effect of the final with examples can be resolved based on the matrix attribute reduction algorithm Based on rough sets and classic attribute reduction algorithm to compare that rough set based on the classic attribute reduction algorithm and the inadequacy of the VMAR algorithm advantage. Key words: rough set; precision rough