基于概念格模型的约简表示及求解1.pdf

基于概念格模型的约简表示及求解1 胡学钢 王昕娅2 合肥工业大学计算机与信息学院 合肥 230009 e-mail ：jsjxhuxg@ crystal0085@163.com 摘要：约简与核是粗糙集理论中的核心问题，也是长期以来研究的热点。现有的约简和核的求 解算法主要是采用区分矩阵法和一些启发式方法，其中区分矩阵法对空间要求较高，而启发式 算法往往无法保证结果的完备性。本文在研究概念格与粗糙集之间关系的基础上，建立了基于 概念格的粗糙集合模型表示，并提出了基于概念格模型求解决策表的约简与核的方法。该算法 首先构建概念格，通过在建格过程中及时动态地消除与约简求解无关的结点，提高算法的效率， 继而扫描概念格判断决策表的一致性，得到核以及所有非约简，进而得到所有约简。理论分析 与实验表明了算法的正确性和有效性。 关键词：核 约简 概念格 粗糙集 The Representation and Resolution of Reduct Based on Concept Lattice Xuegang Hu, Xinya Wang Department of Computer Science and Technology, Hefei University of Technology, Hefei 230009, China E-MAIL: jsjxhuxg@, crystal0085@163.com Abstract: Reduct and core are important concepts in rough set theory, and also the hotspot of research. Today the reduct often obtain by using discernibility matrix or some heuristic algorithms. Discernibility matrix needs large memory, and heuristic algorithms are often incomplete. This paper build the rough set model based on concept lattice, and present the reduct algorithm based on concept lattice by the research of the relation between concept lattice and rough set. Its main idea is: concept lattice was first built, and eliminate unnecessary nodes timely and dynamically in building process to improving the efficiency, then the consistency, all the non-reducts and core of the decision table were obtained by scanning the concept lattice, at last all the reducts were obtained. Theory analysis and the experiment show the algorithm is correct and effective. Key Words: core, reduct, concept lattice, rough set 1.引言