基于差别矩阵的属性集求核算法  

A Core Algorithm of Attribute Sets Based on the Discernibility Matrix

作　　者：杨涛 张贤勇[1,2] 冯山[1] YANG Tao;ZHANG Xianyong;FENG Shan(College of Mathematics and Software Science,Sichuan Normal University,Chengdu 610066,China;Institute of Intelligent Information and Quantum Information,Sichuan Normal University, Chengdu 610066,C)

机构地区：[1]四川师范大学数学与软件科学学院,四川成都610066 [2]四川师范大学智能信息与量子信息研究所,四川成都610066

出　　处：《郑州大学学报（理学版）》2018年第1期27-32,共6页Journal of Zhengzhou University:Natural Science Edition

基　　金：国家自然科学基金项目(61673285;61203285);四川省青年科技基金项目(2017JQ0046);四川省教育厅科研项目(15ZB0029)

摘　　要：通过对差别矩阵的研究,提出一种压缩差别矩阵的构造方法及相关的属性集求核算法,减少了差别矩阵中大量的空值存储,并且避免对象之间的盲目比较.算法的时间复杂度为max(O(|C||U_1||U_2|),O(|C'||U_1'|~2)),空间复杂度为max(O(|C||U_1'||U_2'),O(|U_1'|~2)),提高了已有差别矩阵算法的求核效率.实例分析与实验结果均验证了构建算法的有效性.Through the study of the discernibility matix,a constructional method of the compression discernibility matrix and the core algorithm of attribute sets were proposed.This algorithm reduced the large amount of empty values which were stored in the discernibility matrix,and avoided the blind comparison between objects.The time complexity was max(O(∣C∣∣ U 1∣∣U 2∣),O(∣C′∣∣U′1∣2)),while the space complexity was max(O(∣C∣∣U′1∣∣ U′2∣),O(∣U′1∣ 2)).The new algorithm promoted the core calculation efficiency of existing algorithms based on the discernibility matrix.Both the example analyses and experiment results verified the effectiveness of the constructional algorithm.

关 键 词：粗糙集 差别矩阵 属性集 核 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]