一种新的粗糙集约简界说及其模糊扩展研究.pdfVIP

摘要 摘 要 在粗糙集理论的代数观点下 约简关注的只有数据库中相容示例的部 分 而在粗糙集理论的信息论观点下 约简开始关注库中不相容的部分 试 图保证数据库中所有数据的不可分辨关系不发生变化 但是我们验证了它的 不完全性 并给出一种新的约简的定义 这样定义的约简完全体现了原数据 库中所有数据的不可分辨关系 弥补了信息论的这一不足 通过对约简算法 的研究 提出了一种不一致度增加函数启发的属性约简算法 并从理论与实 验两方面验证了它的正确性及有效性 将建立在分明集上的一些概念及相应 结论推广到模糊集上是一项很有意义的工作 本文将粗糙集理论的核心概 念――约简与核 推广到了模糊集上 定义了――模糊约简与模糊核 这种 扩展研究有希望导致一种新的Rough fuzzy 理论 关键词 粗糙集 约简 约简算法 模糊约简 I Abstract Abstract In rough set theory, the reduction in the algebra view is only concerned about the consistent part in a database. While in the information view, reduction is concerned about the all data, inconsistent part included. Reduction in information view tries to embody the discernibility relation between the all data. Whereas this try is not perfect. So this paper gives a new definition of reduction, which is just to redeem this lack and it can completely embody the discernibility relation. After the research in the reduction algorithm, a new heuristic reduction algorithm based on inconsistency increasing function is proposed. This algorithm proves its validity in theory and in experiment. As we all know, it is significant to generalize some concepts and results defined on a crisp set to counterpart defined on a fuzzy set. We generalize the key concepts — reduction and core defined on a crisp set to their counterpart — fuzzy reduction and fuzzy core. This work is expected to initiate a new fuzzy-rough theory. Key Words: Rough sets; Reduction; Reduction algorithm; Fuzzy reduction 第 1 章 绪论 第1章 绪论 1.1粗糙集理论研究发展的历史 现状及发展趋势分析 粗糙集作为一种处理不精确 不确定与不完全数据的有力的数学工具, 最