摘要
从粗糙集理论出发,在可分辨关系和对象差异矩阵概念的基础上构造出基于粗糙集理论的并行约简算法。算法首先将原系统划分为多个子系统,然后利用评价指数对划分得到的子系统并行求解,最后以子系统的局部约简结果为基础,求得原系统的约简。算法的时空性能较好,适于处理大规模数据集。
A new parallel algorithm based on rough set was proposed after introducing the concepts of distinguishable relation and distinguishable matrix. The algorithm divided the system into several sub-systems, and then the evaluation index was used for the parallel computing of the sub-systems. Finally, the original system＇s reduction was got based on the part reduction results of the sub-system. The proposed algorithm has good performance in spatio-temporal, and it is good at dealing with the huge volume of data.
出处
《计算机应用》
CSCD
北大核心
2007年第8期1964-1966,共3页
journal of Computer Applications
关键词
数据挖掘
粗糙集
属性约简
并行
data mining
rough set
attribute reduction
parallel
