K-均值聚类算法的MapReduce模型实现

针对日益严峻的大数据处理时间长、执行速率低等问题,通过深入分析,提出了一种提高大规模数据聚类效率的方法。以K-均值聚类算法为原型,利用Map Reduce模型在大规模数据处理方面的优势,对原有算法进行并行化改进,设计出一种基于Hadoop分布式云平台的K-均值聚类Map Reduce模型。应用此模型,对淘宝用户仿真数据进行聚类试验,试验结果表明,对K-均值聚类算法的Map Reduce模型实现后,性能优于原算法性能,缩短了聚类时间,提高了聚类效率,特别适于对海量数据进行聚类处理。

The K-means Clustering Algorithm Research Based on the MapReduce Model

Increasingly grim for a long time big data processing, and low execution rate, through in-depth analysis, this paper presents a method to improve the efficiency of large-scale data clustering methods.K- means clustering algo-rithm to prototype, utilizing the advantages of MapReduce model for large-scale data processing, the original algorithm parallelization improvements designed K- means clustering algorithm model based on Hadoop MapReduce distributed cloud platform .Using this model,the simulation data for Taobao users to cluster trial,which demonstrated the feasibili-ty of this method,shortening the clustering time,especially suitable for massive data clustering process.