一种基于MapReduce的粗糙集并行属性约简算法  

A parallel rough set attribute reduction algorithm based on MapReduce

作　　者：杨勇[1] 朱影[1] 

机构地区：[1]重庆邮电大学计算智能重庆市重点实验室,重庆400065

出　　处：《重庆邮电大学学报（自然科学版）》2015年第1期89-96,共8页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：重庆市自然科学基金(CSTC;2007BB2445);重庆市教委科学技术研究项目(KJ110522);重庆邮电大学科研基金(A2009-26)~~

摘　　要：云计算技术是海量数据挖掘的一种高效解决方案,将MapReduce并行计算模型与粗糙集属性约简算法相结合,提出一种基于MapReduce的浓缩布尔矩阵并行属性约简算法。该算法提高了粗糙集属性约简算法对大数据的处理能力和效率,并能适应云计算环境。实验结果表明,所提算法具有良好的效率、加速比和可扩展性。Cloud computing technology is a high efficient solution for massive data mining. MapReduce,the parallel computing model is combined with attribute reduction algorithm of rough set,and a novel parallel concentration Boolean matrix attribute reduction algorithm is proposed in this paper. The algorithm improves the capacity and efficiency of the rough set attribute reduction algorithms for big data. Furthermore,it also adapts to the cloud computing environment. The experimental results illustrate the high efficiency,speedup and scaleup of the proposed algorithm.

关 键 词：MAPREDUCE 粗糙集 属性约简 浓缩布尔矩阵 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]