基于Hadoop平台的DBIK-means聚类算法并行化研究

Research on parallel DBIK- means clustering algorithm based on Hadoop platform

针对Hadoop存在的不足,提出了利用分布式数据库来模拟共享存储空间的解决方案。并对DBIK-means聚类算法做了并行化设计和实验分析,验证了DBIK-means聚类算法在处理大数据时,能够获得较好的加速比。

According to the existing problems of the Hadoop, the scheme of using the distributed database to simulate the sharing storage space is put forward. The clustering algorithm based on DBIK - means is parallel designed and analysed for experiments to verify that the DBIK -means clustering algorithm can get good speedup in processing large data.