【作者】 赵磊；

【导师】 赵明清；

【作者基本信息】 山东科技大学 ， 情报学， 2011， 硕士

【摘要】 粗糙集理论和方法是一种能有效分析和处理不一致、不精确、不完备等各种信息的数据分析工具。该理论和方法已经在模式识别、机器学习、决策支持、知识发现、预测建模等领域得到成功的应用。知识粒度的计算和属性约简作为粗糙集理论和应用的关键技术,已经成为了备受关注的研究热点。本文研究的主要内容包括：介绍了粗糙集理论和粒度理论的国内外研究现状和基本概念,在此基础上对已有的知识粒度形式进行了更为一般化的推广,得到了组合粒度和多项式粒度,它们包含了目前常见的知识粒度,并讨论了这些粒度的性质问题。在信息系统中常用的属性约简算法的基础上给出了一种新的基于粒度的属性重要度的属性约简算法和一种基于粒度意义下的属性约简算法。这些成果对于在信息系统中建立粒度计算以及属性约简有一定的理论意义和应用价值。更多还原

【Abstract】 The theory and method of Rough set is a data analysis tool able to analyze and deal with inconsistent, inaccurate, incomplete information effectively. The theory and methods have been application successful in pattern recognition, machine learning, decision support, knowledge discovery, predictive modeling and other areas. Calculation of knowledge granularity and properties of knowledge reduction as the rough set theory and application of key technologies has become the research focus of concern.The main contents of this paper include:Introduced the basic concepts and research status theory of rough set theory and size theory. Proposed the more generalization form of knowledge granularity on the basis of already exists. Get the portfolio granularity and the polynomial granularity. They include the common knowledge granularity, And discuss the nature of these granularity. Given a new importance based on the properties of granularity attribute reduction algorithm. And a reduction algorithm based on the sense of granularity. These results have certain theoretical significance and application value in establish the granularity in the information system and the attribute reduction.更多还原