Mapreduce下改进Skyline的高效算法

目前基于 MapReduce 的 Skyline 算法随着维度增大会陷入维度灾难，不能高效地解决大数据条件下的计算问题。提出高效算法 MRBPS，利用数据间的互不支配特性，通过一个优化轴点对数据集建立区域标识，在 Map 和 Reduce 阶段优先比较每个点的区域标识，将多维比较简化为一维比较，提高了计算效率，通过系统实验证明：此算法在大数据量时能够明显提高计算效率，与现有算法相比具有高效性和可靠性。

Improved efficient skyline algorithm based on Mapreduce

Existing Mapreduce-based Skyline algorithms is inefficient facing large scale database,to solve this problem,an MapReduce with balanced point skyline( MRBPS)algorithm is proposed,using incomparability of dataset,map points to different regions with a computed balanced point,simplified multi-dimensional comparison to one dimensional comparison,reduce number of tests in Map and Reduce Task. Systematic experiments prove that the algorithm is efficient in large scale database,and more efficient and reliable than existing algorithms.