基于动态区分矩阵的属性约简算法  

Attribute Reduction Algorithm Based on Dynamic Discernibility Matrix

作　　者：孙凌宇[1] 彭宣戈[1] 冷明[1,2] 

机构地区：[1]井冈山大学计算机科学系,吉安343009 [2]上海大学计算机工程与科学学院,上海200072

出　　处：《计算机工程》2008年第24期216-217,220,共3页Computer Engineering

基　　金：科技部国际合作基金资助项目(CB7-2-01);江西省教育规划基金资助项目(04BY453)

摘　　要：在分析基于静态区分矩阵的属性约简算法基础上,提出一种基于动态区分矩阵的属性约简算法。该算法采用2种不同的区分矩阵调整方案,使其能客观及时地反映出当前的约简以及剩余条件属性对信息系统的影响。实验结果表明,该算法不仅能找到信息系统的属性约简,还能有效减少计算属性约简的计算量,提高计算效率。On the basis of analyzing an attribute reduction algorithm based on static discernibility matrix, an attribute reduction algorithm based on dynamic discernibility matrix is proposed. Two different adjustment schemes are used to give the existed reduction and the effect of the extra contidion attribute to information system objectively and timely. Experimental results show this algorithm can not only find attribute reduction of information system, but also reduce the quantity for computing the attribute reduction effectively, and promote the efficiency of the computation.

关 键 词：粗糙集 约简 区分矩阵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]