摘要
随着数据流量的指数增长,大数据成为当前计算设备超乎预期的负担.近年来大数据的降维倍受关注,应用降维技术对原始数据进行核心信息提取,可以大幅降低数据存储空间占用.本文提出一种大数据用户偏好信息全局降维算法,该算法主要面向解决分布式大数据降维中的三类核心问题:大数据聚合、降维算法和分布式计算的架构设计.理论分析表明,所提算法可以有效减少数据存储所需空间、提升数据转换效率,且对计算能力配置的需求有明显降低.
Big data becomes a unpredicted big burden of computing capabilities,along with the exponential growth of data volume.Dimension reduction is interested in recent years because its ability to decrease the volume of big data via extracting core information from raw data.A Global Dimension Reduction towards big data users Preference(GDRP)is proposed in this paper.GDRP is proposed to solve problems involving with big data fusion,dimension reduction and distributed network structure design.Theoretic analysis shows that GDRP could low down the storage requirement for big data and enhance the data converting efficiency,as well as cut down the computing infrastructure setting requirement at the same time.
作者
郑羽洁
ZHENG Yu-jie(Computer Science Department,Guangxi Economic Management Cadre College,Nanning 530007,China)
出处
《广西民族大学学报（自然科学版）》
CAS
2018年第4期70-74,共5页
Journal of Guangxi Minzu University ：Natural Science Edition
基金
广西教育科学"十二五"规划2015年度立项课题(2015C455)
关键词
大数据
数据挖掘
降维
big data
data mining
dimension reduction
