摘要
针对传统的非平衡大数据云存储方法兼容性效果差、数据失真率较高等缺陷,提出一种分布式非平衡大数据兼容性云储存方法。首先根据分布式系统的数据存储原理,分析行式和列式储存的消耗情况,并针对消耗现象应用分类器处理正负两类样列,以获得较好的分类效果和几何均值;测评非平衡数据集分类效果,强化云存储方法兼容性;最后通过构建大数据访问局部性优化体制,实现分布式非平衡大数据兼容性云储存。经仿真验证,所提方法的兼容性优、失真率较低。
Due to low compatibility and high data distortion rate in traditional methods, a method of cloud storage distributed unbalanced big data compatibility was proposed. According to the principle of data storage of distributed systems, the consumption of row-based storage and column-based storage was analyzed. On this basis, the positive and negative samples were processed by classifiers, so as to obtain a better classification effect and geometric mean value. Moreover, the classification effect of unbalanced data sets was evaluated and the compatibility of the cloud storage method was enhanced. Finally, the cloud storage of distributed unbalanced big data compatibility was achieved by constructing the local optimization system of big data access. Simulation results prove that the compatibility of the proposed method is good and the distortion rate is low.
作者
康瑞华
KANG Rui-hua(School of Computer,Hubei University of Technology,Wuhan Hubei 430070,China)
出处
《计算机仿真》
北大核心
2021年第6期339-342,347,共5页
Computer Simulation
基金
湖北省教育厅指导项目(B2015051)。
关键词
分布式
非平衡大数据
兼容性
云储存
Distributed
Unbalanced big data
Compatibility
Cloud storage
