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

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.