论文部分内容阅读
Fuzzy rough sets model is powerful to deal with hybrid uncertain data.All the existing algorithms are serial and only run on a single computer,thus they can only deal with small data sets.In this work,we focus on MapReduce-based parallel computation of fuzzy rough approximation,which is the foundation for all applications in fuzzy rough sets.A key step in fuzzy rough approximation is fuzzy membership degree computation.Since fuzzy relation does not satisfy transitivity,the basic MapReduce framework cannot be applied to solve this problem.We propose a novel MapReduce model for computing fuzzy relation,which solve the problem by adding another data pipe.Experimental results demonstrate that the proposed model is effective for parallel computation of fuzzy rough approximation.