属性关联模型下大数据集群查询仿真  

Simulation of Big Data Cluster Query under Attribute Correlation Model

作　　者：周敏 曾达 杨祥[2] ZHOU Min;ZENG Da;YANG Xiang(College of Big Data and Artificial Intelligence,Nanning College of Technology,Nanning Guangxi 530105,China;Guilin University of Technology,Guilin Guangxi 541006,China)

机构地区：[1]南宁理工学院大数据与人工智能学院,广西南宁530105 [2]桂林理工大学,广西桂林541006

出　　处：《计算机仿真》2024年第3期524-527,537,共5页Computer Simulation

基　　金：2022年度广西高校中青年教师科研基础能力提升项目教育信息化专项项目(2022XXH0018);2023年度广西高等教育本科教学改革工程项目(2023JGB490)。

摘　　要：在数据查询过程中,易受冗余数据、服务器异常、虚拟信息等问题的干扰,导致查询时间长、查询稳定性差等现象产生。为了解决上述问题,提出基于属性关联模型的大数据集群查询算法。采用扩展t-SNE算法对大数据集群中的数据节点做降维处理,避免冗余数据对查询过程产生干扰。将降维后的数据输入到属性关联模型中,实现大数据集群的特征提取,并将提取的特征输入到分布式并行架构中,通过查询负载量的计算完成大数据集群的查询。实验结果表明,所提算法的响应时间短,查询开销小于50Mb,且查询稳定性强。During the data query process,it is susceptible to interference from redundant data,server anomalies,virtual information,and other issues,resulting in long query times and poor query stability.Therefore,a query algorithm for big data cluster based on relation attribute model was put forward.Firstly,the extended t-SNE algorithm was adopted to reduce the dimension of data node in big data cluster,thus avoiding the interference from redundant data in the query process.Secondly,the data after dimensionality reduction were input into the relation attribute model to extract the feature of big data cluster.Meanwhile,the extracted feature was input into a distributed parallel architecture.Finally,the query of big data cluster was completed by calculating the query load.Experimental results show that the proposed algorithm has short response time and strong query stability,and the query overhead is less than 50Mb.

关 键 词：大数据集群降维 特征提取 属性特征 分布式并行架构 负载均衡分配 查询负载量 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]