基于核搜索和二进制可辨识矩阵的属性约简算法  

An attribute reduction algorithm of rough set based on Core Searching and binary discernibility matrix

作　　者：崔建华[1] 褚蕾蕾[1] 常文波[1] 

机构地区：[1]西安交通大学理学院,陕西省西安市710049

出　　处：《微计算机信息》2010年第36期180-181,184,共3页Control & Automation

摘　　要：本文结合粗糙集核搜索算法和二进制可辨识矩阵,提出了一种基于二进制可辨识矩阵的约简算法。与基于可辨识矩阵的经典算法相比,本算法有效地降低了时间复杂度,从而适用于大型数据集;与核搜索算法相比,考虑更全面,得到的结果更精确。理论分析和实验表明,本算法是一种快速、高效的粗糙集属性约简算法。An algorithm is put forward in the article,which is based on Core Searching and binary discernibility matrix. This algorithm greatly reduce time complexity for large data sets compare with traditional algorithm and has a more accurate result compare with Core Searching Algorithm because of comprehensive considering. Theoretical analysis and experiment results show it is an effective attribute reduction algorithm.

关 键 词：粗糙集 二进制可辨识矩阵 属性约简 核 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]