摘要
粗糙集是一种处理不精确和不完全数据的工具,其研究的主要内容之一是对数据进行约简,数据经过约简后能更准确的获取知识,现已证明寻找最小约简是NP难问题.由于基于最初差别矩阵属性约简定义与基于正区域约简定义有差异,因此,设计一种新的区分矩阵与免疫遗传算法相结合的方法,能够实现相容/不相容决策表的属性约简,简化差别矩阵,降低不可区分关系的算法复杂度,通过简化区分函数方法求得核属性,有效提高计算速度,并通过实例证明了算法的有效性.
A rough set is a tool used to deal with inaccurate and incomplete data. One of the main research contents of a rough set is to reduce the data,and the data after reduction can be used to obtain knowledge more accurately. It is proved that searching the minimum reduction is an NP-hard problem. Because of the difference between the attribute reduction definition based on the primary difference matrix and the reduction definition base on positive region,a new method combining discernibility matrix with the immune genetic algorithm be designed in this paper,which can realize the attribute reduction of compatibility or incompatibility decision table. This method firstly simplifies the difference matrix,reduces the algorithm complexity of indiscernibility relation,then obtains the core attribute by the method of reducing discernibility function,improves the computing speed efficiently,and finally proves the validity of the algorithm by examples.
出处
《江苏第二师范学院学报》
2016年第6期1-3,共3页
Journal of Jiangsu Second Normal University
关键词
粗糙集
属性约简
区分矩阵
免疫遗传算法
rough set
attribute reduction
discernibility matrix
immune genetic algorithm