基于Hadoop 平台的用户行为挖掘

摘要

随着互联网发展带来的数据爆炸，使得 Web日志的数据量也越来越大，如何从海量的 Web 日志中挖掘有价值的信息成为了目前研究的热点。本文提出基于 Hadoop 集群框架对 Web 日志进行挖掘。实验结果表明，该集群系统既可以处理海量的 web 日志，同时也能够挖掘出有价值的信息，并证实了利用sqoop在 Hive仓库和传统数据库之间数据迁移的可行性。%The rapid development of Internet brings data explosion,so web log data is becoming bigger and bigger.How to mine valuable information from huge amounts of Web log becomes the focus of present study.This paper presented Web log mining based on the Hadoop cluster framework.Experimental results show that this cluster system can process massive web log data,and can mine valuable information.And it is confirmed to be feasible that the data migrates between Hive warehouse and traditional database by using sqoop.