面向云计算的数据挖掘分类算法研究  

Classification algorithm of data mining based on Cloud computing

作　　者：卢龙[1,2] 王静宇[1] 王超 

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010 [2]中国北方稀土(集团)高科技股份有限公司,内蒙古包头014010 [3]中国移动通信集团山东有限公司莱芜分公司,山东莱芜271100

出　　处：《微型机与应用》2017年第6期7-9,12,共4页Microcomputer & Its Applications

基　　金：国家自然科学基金项目(61662056;61462069);内蒙古自然科学基金(2015MS0622;2016MS0609)

摘　　要：针对传统贝叶斯分类算法在处理海量数据时存在的运行时间长和分类准确率低等问题,在对传统的贝叶斯分类算法和云计算进行了深入研究后,提出了面向云计算环境的基于MapReduce模型的朴素贝叶斯分类算法。该算法实现了朴素贝叶斯分类算法的并行化,实现了大规模数据在云计算环境下的集群中进行贝叶斯分类处理。实验结果证明,该算法具有较高的分类准确率,在运行时间和加速比方面也有很好的效果。Aiming at the great challenge that comes from the large-scale automatic text classification of data mining classification algorithm because of the explosive increase of data,after studying traditional Bayesian classification algorithm,this paper proposed a Bayesian classification algorithm based on MaprRduce model for Cloud computing environment,making use of the Bayesian classification method of Cloud computing,dealing with the large-scale data in the Cloud cluster environment through Bayesian classification algorithm. Compared with single machine＇s handling time,experiments show that this method has greater execution efficiency which makes it suitable to process massive data rapidly in discretization way.

关 键 词：云计算 朴素贝叶斯算法 MAPREDUCE 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]