摘要
针对粗糙集属性约简的结果容易出现局部最优问题,引入差别矩阵,将相对和绝对属性约简统一为差别列表上的集合操作,提出基于蚁群系统的启发式数据约简算法,将蚁群算法的启发信息建立在差别矩阵的核度和必要度上。实验仿真表明本算法可以较好地克服普通约简算法的局部最优问题,说明本算法具有较好约简性能。
To solve the rough set reduction problem,this paper introduced the discernibility table based on the discernibility matrix,and unified absolute and relative reduction into one kind of set operation.Then proposed a heuristic rough set reduction algorithm based on ant colony system.The experiment study shows that the new algorithm overcomes the shortcomings of greed method and proves to possess good reduction performance.
出处
《计算机应用研究》
CSCD
北大核心
2011年第3期941-944,共4页
Application Research of Computers
关键词
数据挖掘
差别矩阵
粗糙集
约简
蚁群系统
Data mining
discernibility matrix
rough set
reduction
ant colony system
