基于云计算的物联网数据挖掘

物联网是一个巨大的、分布广泛的物与物相连的网络,其上产生的海量数据通常是与时间和空间相关的,具有动态、异构、分布的特性,因此对这些数据的挖掘非常困难,而且耗费时间和内存,效率低下。为了解决这些问题,提出了一种基于云计算的物联网数据挖掘系统:将这些海量数据转化为PML数据文件,并存储在HDFS中,同时把挖掘任务分配到多台节点服务器上并行处理。采用Hadoop平台,将关联规则Apriori算法Map/Reduce化,提高了挖掘的效率。另外,采用副本策略将计算向存储迁移,将失效节点的计算迁移到副本数据存储节点就地执行,降低了数据传输的时间,大为提高了挖掘的效率。

The Internet of things Data Mining Based On Cloud Computing

The internet of things is a huge and widely distributed the Internet that things connect things, from that,its massive data is usually time and space related dynamic, heterogeneous, distribution and very complex, These data mining is very difficult, and cost a lot of time and memory, the efficiency is also low. In order to solve these problems, this paper proposes a internet of things data mining system based on cloud computing. Put these massive data into PML data file and store in HDFS. At the same time, mining task allot to many sets of node server parallel processing. The system uses the Hadoop platform,put the association rules algorithm of Apriori into Map/Reduce, improve the efficiency of the data mining. In addition, the strategy will be calculated to copy stored mi- gration, The failure of the calculation of the node will move to copy data storage nodes in situ execution, reduce the data transmission time, greatly improving the mining efficiency.