一种改进的SPRINT算法  

An Improved SPRINT Algorithm

作　　者：白玲玲[1] 韩天鹏[2] BAI Ling-ling;HAN Tian-peng(Office of Academic Affairs,Fuyang Party School of CPC,Fuyang 236034;School of Computer and Information Engineering,Fuyang Normal University,Fuyang 236037,Anhui,China)

机构地区：[1]中共阜阳市委党校教务处,安徽阜阳236034 [2]阜阳师范学院计算机与信息工程学院,安徽阜阳236037

出　　处：《韶关学院学报》2018年第9期20-25,共6页Journal of Shaoguan University

基　　金：阜阳市社科规划课题(FSK2017009);阜阳市党校系统科研课题(FYDXKT201741)

摘　　要：自大数据时代以来,数据密集型计算已经引起了相当大的关注.数据密集型计算环境中的数据挖掘研究仍处于初级阶段.提出一种基于MapReduce编程框架和SPRINT算法的决策树分类算法M-BCBT. M-BCBT继承了MapReduce的优点,使算法更适合数据密集型计算应用.算法的性能根据实例进行分析评估.实验结果表明,MBCBT可以缩短操作时间,提高大数据环境的准确性.Data-intensive computing has attracted considerable attention since the advent of the big data era. Data mining research in data-intensive computing environments is still in its infancy. This paper proposes a decision tree classification algorithm M-BCBT based on MapReduee programming framework and SPRINT algorithm. M-BCBT inherits the advantages of MapReduee, making the algorithm more suitable for data-intensive computing applications. The performance of the algorithm is evaluated based on living examples. Experimental results show that M-BCBT can shorten the operation time and improve the accuracy of big data environment.

关 键 词：SPRINT MAPREDUCE 决策树 数据挖掘 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]